feat(playground): top-p and a thoughts switch (auto/off/on)

Playground gains top-p and a reasoning selector matching the Local-models row:
auto follows the model/row default; off/on override per request via
enable_thinking (cortiq >= 0.2.1 for local models). /admin/api/test(+stream)
accept top_p and enable_thinking, proxied through GenParams passthrough.

Cargo.lock

@@ -285,7 +285,7 @@ checksum = "1d07550c9036bf2ae0c684c4297d503f838287c83c53686d05370d0e139ae570"
 
 [[package]]
 name = "cortiq-gateway"
-version = "0.2.7"
+version = "0.2.23"
 dependencies = [
  "anyhow",
  "arc-swap",

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cortiq-gateway"
-version = "0.2.7"
+version = "0.2.23"
 edition = "2021"
 rust-version = "1.75"
 license = "Apache-2.0"

src/admin/mod.rs

@@ -165,7 +165,7 @@ pub fn api_routes(admin_token: String) -> Router<SharedState> {
             "/admin/api/secrets",
             get(list_secrets).put(set_secret).delete(clear_secret),
         )
-        .route("/admin/api/stats", get(get_stats))
+        .route("/admin/api/stats", get(get_stats).delete(clear_stats))
         .route("/admin/api/shadow", get(get_shadow))
         .route("/admin/api/hf/search", get(hf_search))
         .route("/admin/api/import", get(list_imports).post(start_import))
@@ -178,6 +178,8 @@ pub fn api_routes(admin_token: String) -> Router<SharedState> {
         .route("/admin/api/cmf", get(cmf_status))
         .route("/admin/api/cmf/install", post(cmf_install))
         .route("/admin/api/cmf/port", get(cmf_port_check))
+        .route("/admin/api/cmf/files", get(cmf_files))
+        .route("/admin/api/provider/models", post(provider_models))
         .route("/admin/api/requests", get(get_requests))
         .route("/admin/api/test", post(run_test))
         .route("/admin/api/test/stream", post(run_test_stream))
@@ -390,18 +392,105 @@ async fn router_probe(State(state): State<SharedState>) -> ApiResult {
     ok(json!({ "ok": true, "status": "ok", "latency_ms": latency_ms, "usage": usage }))
 }
 
+#[derive(Deserialize)]
+struct ProviderModelsBody {
+    provider: String,
+    base_url: String,
+    /// Key typed into the form (takes priority)…
+    #[serde(default)]
+    api_key: Option<String>,
+    /// …or the name of a stored secret to reuse if no key was typed.
+    #[serde(default)]
+    api_key_env: Option<String>,
+}
+
+/// Concise error line from a provider's JSON error body.
+fn provider_err(body: &str, status: u16) -> String {
+    if let Ok(v) = serde_json::from_str::<Value>(body) {
+        if let Some(m) = v["error"]["message"]
+            .as_str()
+            .or_else(|| v["error"].as_str())
+            .or_else(|| v["message"].as_str())
+        {
+            return format!("{status}: {m}");
+        }
+    }
+    format!("HTTP {status}")
+}
+
+/// Validate a provider API key and list its models by calling the provider's
+/// `/models` endpoint — powers the "Test key & load models" button. The key is
+/// used transiently (from the form) or resolved from a stored secret; it is not
+/// persisted here.
+async fn provider_models(
+    State(state): State<SharedState>,
+    Json(b): Json<ProviderModelsBody>,
+) -> ApiResult {
+    let base = b.base_url.trim_end_matches('/').to_string();
+    if base.is_empty() {
+        return Err(ApiError::bad("base URL is required"));
+    }
+    let key = b.api_key.filter(|k| !k.trim().is_empty()).or_else(|| {
+        b.api_key_env
+            .as_deref()
+            .and_then(|e| state.secrets.resolve(e))
+    });
+    let key = key.as_deref();
+
+    let client = reqwest::Client::builder()
+        .timeout(std::time::Duration::from_secs(15))
+        .build()
+        .map_err(|e| ApiError::bad(e.to_string()))?;
+
+    // Anthropic uses x-api-key + anthropic-version and /v1/models; every other
+    // (OpenAI-compatible) provider uses a Bearer token and /models.
+    let resp = if b.provider == "anthropic" {
+        let mut r = client
+            .get(format!("{base}/v1/models"))
+            .header("anthropic-version", "2023-06-01");
+        if let Some(k) = key {
+            r = r.header("x-api-key", k);
+        }
+        r.send().await
+    } else {
+        let mut r = client.get(format!("{base}/models"));
+        if let Some(k) = key {
+            r = r.header("authorization", format!("Bearer {k}"));
+        }
+        r.send().await
+    };
+    let resp = resp.map_err(|e| ApiError::bad(format!("request failed: {e}")))?;
+    let status = resp.status().as_u16();
+    let body = resp.text().await.unwrap_or_default();
+    if !(200..300).contains(&status) {
+        return ok(json!({ "ok": false, "status": status, "error": provider_err(&body, status) }));
+    }
+    // OpenAI + Anthropic both return { "data": [ { "id": ... } ] }.
+    let v: Value = serde_json::from_str(&body).unwrap_or(json!({}));
+    let mut models: Vec<String> = v["data"]
+        .as_array()
+        .map(|arr| {
+            arr.iter()
+                .filter_map(|m| m["id"].as_str().map(|s| s.to_string()))
+                .collect()
+        })
+        .unwrap_or_default();
+    models.sort();
+    ok(json!({ "ok": true, "count": models.len(), "models": models }))
+}
+
 async fn meta() -> ApiResult {
     ok(json!({
         "version": env!("CARGO_PKG_VERSION"),
-        "providers": ["openai", "anthropic", "ollama", "http"],
+        "providers": ["openai", "anthropic", "openrouter", "lmstudio", "ollama", "http"],
-        "provider_impl": { "openai": true, "ollama": true, "http": true, "anthropic": true },
+        "provider_impl": { "openai": true, "anthropic": true, "openrouter": true, "lmstudio": true, "ollama": true, "http": true },
-        "cost_tiers": ["cheap", "mid", "expensive"],
+        "cost_tiers": ["cheap", "mid", "expensive", "local"],
         "kinds": ["chat", "embedding"],
         "profiles": ["cost-saver", "balanced", "quality-first"],
         "text_strategies": ["last_user", "last_user_plus_system", "concat_all"],
         "policy_modes": ["fixed_table", "cost_aware"],
         "tiers": ["low", "medium", "high"],
-        "caps": ["tools", "vision"],
+        "caps": ["tools", "vision", "audio", "reasoning", "json", "caching"],
         "protocols_impl": {
             "openai_chat": true, "openai_completions": true, "openai_embeddings": true,
             "openai_models": true, "anthropic_messages": true, "mcp": true, "native_passthrough": true
@@ -441,20 +530,20 @@ async fn list_models(State(state): State<SharedState>) -> ApiResult {
         })
         .collect();
 
-    // The managed local CMF server is registered in the routing registry from
+    // Managed local CMF servers are registered in the routing registry from the
-    // the [cmf] flags, not from [[models]]. Surface it here (marked managed, so
+    // [cmf] config, not from [[models]]. Surface each here (marked managed, so
-    // the UI shows it read-only) so the real running model is visible and
+    // the UI shows it read-only) so the real running models are visible and
-    // pinnable without needing a hand-maintained static pool entry. A fresh
+    // pinnable without a hand-maintained static pool entry. A fresh install with
-    // install with no managed model still returns an empty list.
+    // no managed model still returns an empty list.
     let cmf = &live.cfg.cmf;
-    if cmf.manage_server
+    for s in cmf.effective_servers() {
-        && !cmf.local_model.trim().is_empty()
+        if models.iter().any(|m| m["id"] == json!(s.id)) {
-        && !models.iter().any(|m| m["id"] == json!(cmf.model_id))
+            continue;
-    {
+        }
         models.push(json!({
-            "id": cmf.model_id,
+            "id": s.id,
             "provider": "openai",
-            "base_url": format!("http://{}:{}/v1", cmf.local_host, cmf.local_port),
+            "base_url": format!("http://{}:{}/v1", cmf.local_host, s.port),
             "model": "cortiq",
             "cost_tier": "local",
             "price_in": 0.0,
@@ -517,7 +606,10 @@ fn cmf_file_for_model(
     cmf: &crate::config::CmfCfg,
     m: &crate::config::ModelCfg,
 ) -> Option<std::path::PathBuf> {
+    // Any loopback URL counts as local — managed servers may sit on any port.
     let local = m.base_url == cmf.cortiq_server_url
+        || m.base_url.contains("//127.0.0.1:")
+        || m.base_url.contains("//localhost:")
         || m.base_url
             .contains(&format!("{}:{}", cmf.local_host, cmf.local_port));
     let stem = m.model.trim();
@@ -544,25 +636,59 @@ fn strip_id_from_routing(routing: &mut RoutingCfg, id: &str) {
     }
 }
 
-async fn delete_model(State(state): State<SharedState>, Path(id): Path<String>) -> ApiResult {
+/// Whether any managed server or static model in `cfg` still uses `path` as its
-    let mut cfg = current_cfg(&state);
+/// local `.cmf`. Guards against deleting a file that another model shares.
+fn cmf_path_referenced(cfg: &crate::config::Config, path: &std::path::Path) -> bool {
+    cfg.cmf
+        .effective_servers()
+        .iter()
+        .any(|s| std::path::Path::new(&s.model) == path)
+        || cfg
+            .models
+            .iter()
+            .any(|m| cmf_file_for_model(&cfg.cmf, m).as_deref() == Some(path))
+}
 
-    // The managed local model lives in [cmf], not [[models]]. Deleting it stops
+#[derive(Deserialize)]
-    // the running server, clears the managed config, and removes its file.
+struct DeleteModelQuery {
-    if id == cfg.cmf.model_id
+    /// Also remove the local `.cmf` from disk. Defaults to true (the historical
-        && cfg.cmf.manage_server
+    /// behaviour); the UI asks the user and passes it explicitly.
-        && !cfg.cmf.local_model.trim().is_empty()
+    delete_file: Option<bool>,
-        && !cfg.models.iter().any(|x| x.id == id)
+}
-    {
+
-        let file = cfg.cmf.local_model.clone();
+async fn delete_model(
-        cfg.cmf.local_model = String::new();
+    State(state): State<SharedState>,
+    Path(id): Path<String>,
+    Query(q): Query<DeleteModelQuery>,
+) -> ApiResult {
+    let mut cfg = current_cfg(&state);
+    let delete_file = q.delete_file.unwrap_or(true);
+
+    // Managed local models live in [cmf], not [[models]]. Deleting one stops its
+    // server, removes it from the managed config, and (when asked) deletes its
+    // file — unless another model still points at the same file.
+    let managed = cfg.cmf.effective_servers().into_iter().find(|s| s.id == id);
+    if let Some(server) = managed {
+        if !cfg.models.iter().any(|x| x.id == id) {
+            if cfg.cmf.servers.iter().any(|s| s.id == id) {
+                cfg.cmf.servers.retain(|s| s.id != id); // multi-model list entry
+            } else {
+                cfg.cmf.local_model = String::new(); // legacy single model
                 cfg.cmf.manage_server = false;
+            }
             strip_id_from_routing(&mut cfg.routing, &id);
-        state.cmf.stop().await;
+            let path = std::path::PathBuf::from(&server.model);
+            let shared = cmf_path_referenced(&cfg, &path);
+            state.cmf.stop_one(&id).await;
             state.reload(cfg)?;
-        let file_removed = remove_cmf_path(std::path::Path::new(&file));
+            let file_removed = if delete_file && !shared {
+                remove_cmf_path(&path)
+            } else {
+                false
+            };
             return ok(json!({ "ok": true, "file_removed": file_removed }));
         }
+    }
 
     let victim = cfg.models.iter().find(|x| x.id == id).cloned();
     let before = cfg.models.len();
@@ -574,23 +700,34 @@ async fn delete_model(State(state): State<SharedState>, Path(id): Path<String>)
     let file = victim
         .as_ref()
         .and_then(|m| cmf_file_for_model(&cfg.cmf, m));
+    // Only remove the file if asked to AND no remaining model still references it.
+    let file = file.filter(|p| delete_file && !cmf_path_referenced(&cfg, p));
     state.reload(cfg)?;
-    // Physically remove the local .cmf after the config is persisted.
     let file_removed = file.map(|p| remove_cmf_path(&p)).unwrap_or(false);
     ok(json!({ "ok": true, "file_removed": file_removed }))
 }
 
 async fn probe_model(State(state): State<SharedState>, Path(id): Path<String>) -> ApiResult {
     let live = state.live();
-    let mcfg = live
+    // Provider string: from the static pool, or "openai" for a managed local
+    // model (which is synthesized into the registry, not [[models]]).
+    let provider_str = live
         .cfg
         .models
         .iter()
         .find(|m| m.id == id)
-        .ok_or_else(|| ApiError::not_found(format!("model '{id}' not found")))?
+        .map(|m| m.provider.clone())
-        .clone();
+        .or_else(|| {
+            live.cfg
+                .cmf
+                .effective_servers()
+                .iter()
+                .find(|s| s.id == id)
+                .map(|_| "openai".to_string())
+        })
+        .ok_or_else(|| ApiError::not_found(format!("model '{id}' not found")))?;
 
-    if mcfg.provider == "anthropic" {
+    if provider_str == "anthropic" {
         return ok(json!({
             "ok": false,
             "error": "anthropic provider not yet implemented (planned v0.2)",
@@ -627,7 +764,21 @@ async fn probe_model(State(state): State<SharedState>, Path(id): Path<String>) -
     let latency_ms = started.elapsed().as_millis() as u64;
     match res {
         Ok(_) => ok(json!({ "ok": true, "latency_ms": latency_ms })),
-        Err(e) => ok(json!({ "ok": false, "latency_ms": latency_ms, "error": e.to_string() })),
+        Err(e) => {
+            let msg = e.to_string();
+            // Make auth failures actionable — a raw "User not found" (OpenRouter's
+            // 401) usually means the API key is wrong or the wrong type.
+            let hint = if msg.contains("401")
+                || msg.contains("Unauthorized")
+                || msg.contains("User not found")
+                || msg.contains("No auth")
+            {
+                " — check the API key (for OpenRouter use an inference key, not a provisioning/management key)"
+            } else {
+                ""
+            };
+            ok(json!({ "ok": false, "latency_ms": latency_ms, "error": format!("{msg}{hint}") }))
+        }
     }
 }
 
@@ -757,21 +908,101 @@ async fn put_settings(State(state): State<SharedState>, Json(b): Json<SettingsBo
         needs_restart = true;
         cfg.cache = v;
     }
+    let mut cmf_apply: Option<(Vec<crate::config::CmfServer>, CmfCfg)> = None;
     if let Some(v) = b.cmf {
-        // manage_server / local_model / threads / gpu apply on restart (the server
+        let old_servers = cfg.cmf.effective_servers();
-        // is spawned once at startup with those env vars); local_only / router
-        // routing apply immediately.
-        if v.manage_server != cfg.cmf.manage_server
-            || v.local_model != cfg.cmf.local_model
-            || v.threads != cfg.cmf.threads
-            || v.gpu != cfg.cmf.gpu
-        {
-            needs_restart = true;
-        }
         cfg.cmf = v;
+        // A managed row supersedes a same-id STATIC pool entry that points at a
+        // local server (the shape Import→register creates): drop the static one
+        // instead of failing validation with "duplicate managed model id".
+        let managed_ids: std::collections::HashSet<String> = cfg
+            .cmf
+            .effective_servers()
+            .into_iter()
+            .map(|s| s.id)
+            .collect();
+        cfg.models.retain(|m| {
+            !(managed_ids.contains(&m.id)
+                && (m.base_url.contains("//127.0.0.1:") || m.base_url.contains("//localhost:")))
+        });
+        // Replacing or renaming a managed model must not brick the save: routing
+        // may still reference the old id. Map a vanished id to its replacement
+        // (the new server on the same port), strip ids that are truly gone, and
+        // repair the default so validate() passes.
+        heal_routing_after_cmf_change(&mut cfg, &old_servers);
+        // Server-list changes apply LIVE after a successful reload: removed /
+        // redefined servers are stopped (RAM freed), new ones are spawned.
+        if cfg.cmf.effective_servers() != old_servers {
+            cmf_apply = Some((old_servers, cfg.cmf.clone()));
+        }
     }
     state.reload(cfg)?;
-    ok(json!({ "ok": true, "needs_restart": needs_restart }))
+    let applied_live = cmf_apply.is_some();
+    if let Some((old, new_cfg)) = cmf_apply {
+        tokio::spawn(crate::cmf_runtime::apply_server_changes(
+            state.cmf.clone(),
+            old,
+            new_cfg,
+        ));
+    }
+    ok(json!({ "ok": true, "needs_restart": needs_restart, "applied_live": applied_live }))
+}
+
+/// After the managed-servers list changed, rewrite [routing] so it only
+/// references ids that still exist. A removed server whose port is now held by
+/// a NEW server is treated as a replacement (old id → new id); other vanished
+/// ids are stripped. An emptied default falls back to the first managed server,
+/// then the first pool model.
+fn heal_routing_after_cmf_change(cfg: &mut Config, old_servers: &[crate::config::CmfServer]) {
+    let servers = cfg.cmf.effective_servers();
+    let known: std::collections::HashSet<String> = cfg
+        .models
+        .iter()
+        .map(|m| m.id.clone())
+        .chain(servers.iter().map(|s| s.id.clone()))
+        .collect();
+    let new_by_port: HashMap<u16, String> =
+        servers.iter().map(|s| (s.port, s.id.clone())).collect();
+    let mut rename: HashMap<String, String> = HashMap::new();
+    for old in old_servers {
+        if !known.contains(&old.id) {
+            if let Some(new_id) = new_by_port.get(&old.port) {
+                rename.insert(old.id.clone(), new_id.clone());
+            }
+        }
+    }
+    let fix = |id: &str| -> Option<String> {
+        if known.contains(id) {
+            Some(id.to_string())
+        } else {
+            rename.get(id).cloned()
+        }
+    };
+    if !cfg.routing.default.is_empty() {
+        cfg.routing.default = fix(&cfg.routing.default).unwrap_or_default();
+    }
+    if cfg.routing.default.is_empty() {
+        cfg.routing.default = servers
+            .first()
+            .map(|s| s.id.clone())
+            .or_else(|| cfg.models.first().map(|m| m.id.clone()))
+            .unwrap_or_default();
+    }
+    for tt in cfg.routing.tiers.values_mut() {
+        let list = match tt {
+            TierTargets::List(v) => v.clone(),
+            TierTargets::One(s) => vec![s.clone()],
+        };
+        let mut healed: Vec<String> = Vec::new();
+        for id in &list {
+            if let Some(new_id) = fix(id) {
+                if !healed.contains(&new_id) {
+                    healed.push(new_id);
+                }
+            }
+        }
+        *tt = TierTargets::List(healed);
+    }
 }
 
 // ───────────────────────────── cmf runtime ─────────────────────────────────
@@ -782,19 +1013,33 @@ async fn cmf_status(State(state): State<SharedState>) -> ApiResult {
     let cmf = state.live().cfg.cmf.clone();
     let ver = crate::cmf_runtime::installed_version(&cmf.cortiq_bin);
     let st = state.cmf.status();
+    // Each configured managed server, merged with its live runtime status.
+    let servers: Vec<Value> = cmf
+        .effective_servers()
+        .iter()
+        .map(|s| {
+            let live = st.servers.iter().find(|x| x.id == s.id);
+            json!({
+                "id": s.id,
+                "model": s.model,
+                "port": s.port,
+                "threads": s.threads,
+                "gpu": s.gpu,
+                "o1": s.o1,
+                "running": live.map(|l| l.running).unwrap_or(false),
+                "healthy": live.map(|l| l.healthy).unwrap_or(false),
+                "last_error": live.and_then(|l| l.last_error.clone()),
+            })
+        })
+        .collect();
     ok(json!({
         "installed": ver.is_some(),
         "version": ver,
-        "running": st.running,
+        "latest": st.latest_version,
-        "healthy": st.healthy,
-        "last_error": st.last_error,
-        "log": st.log,
-        // serving info so the card can show "serving <model> on :<port>"
         "manage_server": cmf.manage_server,
-        "local_model": cmf.local_model,
-        "model_id": cmf.model_id,
         "local_host": cmf.local_host,
-        "local_port": cmf.local_port,
+        "servers": servers,
+        "log": st.log,
     }))
 }
 
@@ -806,6 +1051,36 @@ async fn cmf_install(State(state): State<SharedState>) -> ApiResult {
     ok(json!({ "started": true }))
 }
 
+/// List the `.cmf` model files available under `models_dir` so the UI can offer
+/// them as a dropdown instead of a hand-typed path. Includes tensor-dir `.cmf`.
+async fn cmf_files(State(state): State<SharedState>) -> ApiResult {
+    let dir = state.live().cfg.cmf.models_dir.clone();
+    let mut entries: Vec<(String, String, Option<u64>, bool)> = Vec::new();
+    if let Ok(rd) = std::fs::read_dir(&dir) {
+        for e in rd.flatten() {
+            let name = e.file_name().to_string_lossy().to_string();
+            if !name.ends_with(".cmf") {
+                continue;
+            }
+            let p = e.path();
+            let is_dir = p.is_dir();
+            let size = if is_dir {
+                None
+            } else {
+                std::fs::metadata(&p).ok().map(|m| m.len())
+            };
+            let path = format!("{}/{}", dir.trim_end_matches('/'), name);
+            entries.push((name, path, size, is_dir));
+        }
+    }
+    entries.sort_by(|a, b| a.0.cmp(&b.0));
+    let files: Vec<Value> = entries
+        .into_iter()
+        .map(|(name, path, size, is_dir)| json!({ "name": name, "path": path, "size": size, "is_dir": is_dir }))
+        .collect();
+    ok(json!({ "dir": dir, "files": files }))
+}
+
 #[derive(Deserialize)]
 struct PortQuery {
     port: u16,
@@ -825,7 +1100,7 @@ async fn cmf_port_check(State(state): State<SharedState>, Query(q): Query<PortQu
     };
     let st = state.cmf.status();
     let port = q.port;
-    let running_here = port == cmf.local_port && st.running;
+    let running_here = st.servers.iter().any(|s| s.port == port && s.running);
 
     // The bind+connect probe is blocking std::net; run it off the async runtime.
     let (available, detail, suggested) = tokio::task::spawn_blocking(move || {
@@ -848,6 +1123,9 @@ async fn cmf_port_check(State(state): State<SharedState>, Query(q): Query<PortQu
     ok(json!({
         "port": port,
         "available": available,
+        // true = held by this gateway's own managed server (a replaced/edited
+        // row frees it on restart) — the UI shows this as info, not an error
+        "managed": running_here,
         "detail": detail,
         "suggested": suggested,
     }))
@@ -1017,6 +1295,13 @@ async fn get_stats(State(state): State<SharedState>, Query(q): Query<StatsQuery>
     ok(snap)
 }
 
+/// Clear all recorded stats/logs (in-memory aggregates, series, recent requests)
+/// and truncate the JSONL log — the dashboard "Clear logs" button.
+async fn clear_stats(State(state): State<SharedState>) -> ApiResult {
+    state.stats.clear();
+    ok(json!({ "ok": true }))
+}
+
 /// Self-warming shadow loop: per-task-type promotion state, Wilson-LB of the
 /// judged pass-rate, and whether the local model is serving it yet.
 async fn get_shadow(State(state): State<SharedState>) -> ApiResult {
@@ -1162,24 +1447,44 @@ struct TestBody {
     temperature: Option<f32>,
     #[serde(default)]
     max_tokens: Option<u32>,
+    #[serde(default)]
+    top_p: Option<f32>,
+    /// Reasoning switch (cortiq ≥ 0.2.1): false = answer without the thinking
+    /// phase; true = force it on. Absent = model/row default.
+    #[serde(default)]
+    enable_thinking: Option<bool>,
 }
 fn default_test_model() -> String {
     "cortiq-auto".into()
 }
 
+/// GenParams for a playground request: native sampling fields + the reasoning
+/// switch proxied via passthrough (providers forward unknown fields as-is).
+fn test_gen_params(b: &TestBody) -> GenParams {
+    let mut params = GenParams {
+        temperature: b.temperature,
+        max_tokens: b.max_tokens,
+        top_p: b.top_p,
+        ..Default::default()
+    };
+    if let Some(v) = b.enable_thinking {
+        params
+            .passthrough
+            .insert("enable_thinking".into(), serde_json::Value::Bool(v));
+    }
+    params
+}
+
 async fn run_test(State(state): State<SharedState>, Json(b): Json<TestBody>) -> ApiResult {
     if b.messages.is_empty() {
         return Err(ApiError::bad("messages must not be empty"));
     }
+    let params = test_gen_params(&b);
     let req = ChatRequest {
         routing: parse_routing(&b.model),
         messages: b.messages,
         tools: vec![],
-        params: GenParams {
+        params,
-            temperature: b.temperature,
-            max_tokens: b.max_tokens,
-            ..Default::default()
-        },
         stream: false,
         meta: RequestMeta {
             account: "playground".into(),
@@ -1228,15 +1533,12 @@ async fn run_test_stream(State(state): State<SharedState>, Json(b): Json<TestBod
     if b.messages.is_empty() {
         return ApiError::bad("messages must not be empty").into_response();
     }
+    let params = test_gen_params(&b);
     let req = ChatRequest {
         routing: parse_routing(&b.model),
         messages: b.messages,
         tools: vec![],
-        params: GenParams {
+        params,
-            temperature: b.temperature,
-            max_tokens: b.max_tokens,
-            ..Default::default()
-        },
         stream: true,
         meta: RequestMeta {
             account: "playground".into(),

src/cmf_runtime.rs

@@ -1,51 +1,68 @@
 //! Local CMF model server lifecycle.
 //!
 //! Closes the loop with <https://github.com/infosave2007/cmf>: install the
-//! `cortiq` CLI from crates.io, keep it up to date, and run `cortiq serve` as a
+//! `cortiq` CLI from crates.io, keep it up to date, and run one `cortiq serve`
-//! managed child process. The served model exposes an OpenAI-compatible API, so
+//! per configured local model as a managed child process. Each served model
-//! the registry registers it as an ordinary provider and the gateway routes to
+//! exposes an OpenAI-compatible API, so the registry registers it as an ordinary
-//! it like any other backend.
+//! provider and the gateway routes to it like any other backend.
 
-use crate::config::CmfCfg;
+use crate::config::{CmfCfg, CmfServer};
 use std::process::Stdio;
 use std::sync::{Arc, Mutex};
 use std::time::Duration;
 use tokio::process::{Child, Command};
 
-/// Observable state of the managed local CMF server (surfaced to the admin API).
+/// Observable state of one managed local model server.
+#[derive(Clone, Debug, Default, serde::Serialize)]
+pub struct CmfServerStatus {
+    /// Pool id it is registered under.
+    pub id: String,
+    /// Path of the `.cmf` being served.
+    pub model: String,
+    /// Port the server binds to.
+    pub port: u16,
+    /// OpenAI base URL the local server is registered under.
+    pub base_url: String,
+    /// The child `cortiq serve` process has been spawned.
+    pub running: bool,
+    /// `/healthz` on the local server answered 200.
+    pub healthy: bool,
+    /// Last fatal error for this server, if any.
+    pub last_error: Option<String>,
+}
+
+/// Observable state of the managed local CMF layer (surfaced to the admin API).
 #[derive(Clone, Debug, Default, serde::Serialize)]
 pub struct CmfStatus {
-    /// `manage_server` is on and a `local_model` is configured.
+    /// `manage_server` is on and at least one local model is configured.
     pub enabled: bool,
     /// Version reported by `cortiq --version` (None = not installed).
     pub installed_version: Option<String>,
     /// Newest `cortiq-cli` on crates.io (populated when `auto_update` runs).
     pub latest_version: Option<String>,
-    /// The child `cortiq serve` process has been spawned.
+    /// Per-model server status.
-    pub running: bool,
+    pub servers: Vec<CmfServerStatus>,
-    /// `/healthz` on the local server answered 200.
-    pub healthy: bool,
-    /// OpenAI base URL the local server is registered under.
-    pub base_url: String,
-    /// Path of the `.cmf` being served.
-    pub model: String,
-    /// Last fatal error, if any.
-    pub last_error: Option<String>,
     /// Recent lifecycle log lines (bounded).
     pub log: Vec<String>,
 }
 
-/// Holds the managed child process and its live status.
+/// A spawned server: its pool id and the child process handle.
+struct Managed {
+    id: String,
+    child: Child,
+}
+
+/// Holds the managed child processes and their live status.
 pub struct CmfRuntime {
     status: Mutex<CmfStatus>,
-    child: Mutex<Option<Child>>,
+    children: Mutex<Vec<Managed>>,
 }
 
 impl CmfRuntime {
     pub fn new() -> Arc<Self> {
         Arc::new(Self {
             status: Mutex::new(CmfStatus::default()),
-            child: Mutex::new(None),
+            children: Mutex::new(Vec::new()),
         })
     }
 
@@ -59,32 +76,140 @@ impl CmfRuntime {
         let mut s = self.status.lock().unwrap();
         s.log.push(msg);
         let n = s.log.len();
-        if n > 50 {
+        if n > 80 {
-            s.log.drain(0..n - 50);
+            s.log.drain(0..n - 80);
         }
     }
 
+    /// A global (install/update) failure — not tied to a single server.
     fn fail(&self, msg: impl Into<String>) {
         let msg = msg.into();
         tracing::warn!(target: "cmf", "{msg}");
+        self.status.lock().unwrap().log.push(format!("✗ {msg}"));
+    }
+
+    /// Record a fatal error against a specific server (and the log).
+    fn fail_server(&self, id: &str, msg: impl Into<String>) {
+        let msg = msg.into();
+        tracing::warn!(target: "cmf", "[{id}] {msg}");
         let mut s = self.status.lock().unwrap();
-        s.last_error = Some(msg.clone());
+        if let Some(sv) = s.servers.iter_mut().find(|x| x.id == id) {
-        s.log.push(format!("✗ {msg}"));
+            sv.last_error = Some(msg.clone());
+        }
+        s.log.push(format!("✗ [{id}] {msg}"));
     }
 
-    fn set_child(&self, child: Child) {
+    /// Seed the per-server status slots from the effective server list.
-        *self.child.lock().unwrap() = Some(child);
+    fn init_servers(&self, servers: &[CmfServer], host: &str) {
-        self.status.lock().unwrap().running = true;
+        let mut s = self.status.lock().unwrap();
+        s.enabled = true;
+        s.servers = servers
+            .iter()
+            .map(|sv| CmfServerStatus {
+                id: sv.id.clone(),
+                model: sv.model.clone(),
+                port: sv.port,
+                base_url: format!("http://{host}:{}/v1", sv.port),
+                running: false,
+                healthy: false,
+                last_error: None,
+            })
+            .collect();
     }
 
-    /// Stop the managed server (used on shutdown / reconfigure).
+    fn set_running(&self, id: &str, running: bool) {
+        if let Some(sv) = self
+            .status
+            .lock()
+            .unwrap()
+            .servers
+            .iter_mut()
+            .find(|x| x.id == id)
+        {
+            sv.running = running;
+        }
+    }
+
+    fn set_healthy(&self, id: &str, healthy: bool) {
+        if let Some(sv) = self
+            .status
+            .lock()
+            .unwrap()
+            .servers
+            .iter_mut()
+            .find(|x| x.id == id)
+        {
+            sv.healthy = healthy;
+        }
+    }
+
+    fn add_child(&self, id: String, child: Child) {
+        self.children.lock().unwrap().push(Managed {
+            id: id.clone(),
+            child,
+        });
+        self.set_running(&id, true);
+    }
+
+    fn is_running(&self, id: &str) -> bool {
+        self.status
+            .lock()
+            .unwrap()
+            .servers
+            .iter()
+            .any(|s| s.id == id && s.running)
+    }
+
+    /// Stop all managed servers (used on shutdown / reconfigure).
     pub async fn stop(&self) {
-        if let Some(mut child) = self.child.lock().unwrap().take() {
+        let mut kids = std::mem::take(&mut *self.children.lock().unwrap());
-            let _ = child.start_kill();
+        for m in kids.iter_mut() {
+            let _ = m.child.start_kill();
         }
         let mut s = self.status.lock().unwrap();
-        s.running = false;
+        for sv in s.servers.iter_mut() {
-        s.healthy = false;
+            sv.running = false;
+            sv.healthy = false;
+        }
+    }
+
+    /// Ensure a status slot exists for `server` (used when servers are added at
+    /// runtime); refreshes model/port if the slot already exists.
+    fn add_server_slot(&self, server: &CmfServer, host: &str) {
+        let mut s = self.status.lock().unwrap();
+        s.enabled = true;
+        if let Some(sv) = s.servers.iter_mut().find(|x| x.id == server.id) {
+            sv.model = server.model.clone();
+            sv.port = server.port;
+            sv.base_url = format!("http://{host}:{}/v1", server.port);
+            sv.running = false;
+            sv.healthy = false;
+            sv.last_error = None;
+        } else {
+            s.servers.push(CmfServerStatus {
+                id: server.id.clone(),
+                model: server.model.clone(),
+                port: server.port,
+                base_url: format!("http://{host}:{}/v1", server.port),
+                running: false,
+                healthy: false,
+                last_error: None,
+            });
+        }
+    }
+
+    /// Stop and forget a single managed server (used when a model is deleted).
+    pub async fn stop_one(&self, id: &str) {
+        let removed = {
+            let mut kids = self.children.lock().unwrap();
+            kids.iter()
+                .position(|m| m.id == id)
+                .map(|pos| kids.remove(pos))
+        };
+        if let Some(mut m) = removed {
+            let _ = m.child.start_kill();
+        }
+        self.status.lock().unwrap().servers.retain(|sv| sv.id != id);
     }
 }
 
@@ -92,7 +217,7 @@ impl Default for CmfRuntime {
     fn default() -> Self {
         Self {
             status: Mutex::new(CmfStatus::default()),
-            child: Mutex::new(None),
+            children: Mutex::new(Vec::new()),
         }
     }
 }
@@ -138,7 +263,7 @@ fn parse_ver(s: &str) -> (u32, u32, u32) {
     )
 }
 
-fn version_lt(a: &str, b: &str) -> bool {
+pub(crate) fn version_lt(a: &str, b: &str) -> bool {
     parse_ver(a) < parse_ver(b)
 }
 
@@ -175,26 +300,31 @@ pub async fn install_now(rt: Arc<CmfRuntime>, bin: String) {
     }
 }
 
-/// Spawn `cortiq serve <model> --host <host> --port <port>`; the child is killed
+/// Spawn `cortiq serve <model> --host <host> --port <port>` for one server; the
-/// when its handle is dropped (i.e. when the gateway shuts down).
+/// child is killed when its handle is dropped (i.e. when the gateway shuts down).
-fn spawn_serve(cfg: &CmfCfg) -> Result<Child, String> {
+fn spawn_serve(server: &CmfServer, cfg: &CmfCfg) -> Result<Child, String> {
     let mut cmd = Command::new(&cfg.cortiq_bin);
     cmd.arg("serve")
-        .arg(&cfg.local_model)
+        .arg(&server.model)
         .arg("--host")
         .arg(&cfg.local_host)
         .arg("--port")
-        .arg(cfg.local_port.to_string());
+        .arg(server.port.to_string());
-    // Parallelise decode matvecs across cores (CMF_THREADS). 0 = leave the
+    // Parallelise decode matvecs across cores (CMF_THREADS). 0/1 = leave the
     // runtime's own default; N>1 gave ~2.8x on a 15B CPU model.
-    if cfg.threads > 1 {
+    if server.threads > 1 {
-        cmd.env("CMF_THREADS", cfg.threads.to_string());
+        cmd.env("CMF_THREADS", server.threads.to_string());
     }
     // Offload large matvecs to the GPU (Metal on macOS). Memory-bandwidth-bound
     // on unified memory, so mostly a prefill win; opt-in.
-    if cfg.gpu {
+    if server.gpu {
         cmd.env("CMF_GPU", "1");
     }
+    // Serve-time O(1) attention override: "off" forces exact attention (undoes a
+    // too-aggressive --o1 conversion hint without re-converting the file).
+    if !server.o1.trim().is_empty() {
+        cmd.env("CMF_O1", server.o1.trim());
+    }
     cmd.stdout(Stdio::null())
         .stderr(Stdio::null())
         .kill_on_drop(true)
@@ -222,24 +352,73 @@ async fn wait_healthy(host: &str, port: u16, tries: u32) -> bool {
     false
 }
 
-/// Orchestrate the whole lifecycle: ensure installed → maybe update → spawn →
+/// Apply a changed managed-servers list LIVE (no gateway restart): stop servers
-/// wait healthy. Safe to call in a background task; it never panics.
+/// that were removed or redefined — freeing their memory immediately — and
-pub async fn manage(rt: Arc<CmfRuntime>, cfg: CmfCfg) {
+/// spawn servers that are new or redefined.
-    if !cfg.manage_server || cfg.local_model.trim().is_empty() {
+pub async fn apply_server_changes(rt: Arc<CmfRuntime>, old: Vec<CmfServer>, cfg: CmfCfg) {
+    let new = cfg.effective_servers();
+    // 1. Stop what's gone or changed (kills the child → RAM released).
+    for o in &old {
+        let unchanged = new.iter().any(|n| n == o);
+        if !unchanged {
+            rt.stop_one(&o.id).await;
+            rt.log(format!("stopped '{}' (removed or changed)", o.id));
+        }
+    }
+    if new.is_empty() {
         return;
     }
-    {
+    // 2. Spawn what's new or changed.
-        let mut s = rt.status.lock().unwrap();
+    if installed_version(&cfg.cortiq_bin).is_none() {
-        s.enabled = true;
+        rt.fail("cortiq binary not found — cannot start local model servers");
-        s.base_url = format!("http://{}:{}/v1", cfg.local_host, cfg.local_port);
-        s.model = cfg.local_model.clone();
-    }
-    if !std::path::Path::new(&cfg.local_model).exists() {
-        rt.fail(format!("local_model not found: {}", cfg.local_model));
         return;
     }
+    for n in &new {
+        let unchanged = old.iter().any(|o| o == n);
+        if unchanged {
+            continue; // already running with the same definition
+        }
+        rt.add_server_slot(n, &cfg.local_host);
+        if !std::path::Path::new(&n.model).exists() {
+            rt.fail_server(&n.id, format!("model file not found: {}", n.model));
+            continue;
+        }
+        rt.log(format!(
+            "starting: cortiq serve {} --host {} --port {}",
+            n.model, cfg.local_host, n.port
+        ));
+        match spawn_serve(n, &cfg) {
+            Ok(child) => rt.add_child(n.id.clone(), child),
+            Err(e) => {
+                rt.fail_server(&n.id, format!("failed to spawn: {e}"));
+                continue;
+            }
+        }
+        let (rtc, host, id, port) = (rt.clone(), cfg.local_host.clone(), n.id.clone(), n.port);
+        tokio::spawn(async move {
+            if wait_healthy(&host, port, 240).await {
+                rtc.set_healthy(&id, true);
+                rtc.log(format!(
+                    "local CMF server ready — '{id}' at http://{host}:{port}/v1"
+                ));
+            } else {
+                rtc.fail_server(&id, "did not become healthy in time");
+            }
+        });
+    }
+}
 
-    // 1. Ensure the `cortiq` binary is available.
+/// Orchestrate the whole lifecycle: ensure installed → maybe update → spawn each
+/// configured local model → wait healthy. Safe to call in a background task; it
+/// never panics.
+pub async fn manage(rt: Arc<CmfRuntime>, cfg: CmfCfg) {
+    let servers = cfg.effective_servers();
+    if servers.is_empty() {
+        return;
+    }
+    rt.init_servers(&servers, &cfg.local_host);
+
+    // 1. Ensure the `cortiq` binary is available (once for all servers).
     let mut ver = installed_version(&cfg.cortiq_bin);
     if ver.is_none() {
         if cfg.auto_install {
@@ -285,27 +464,38 @@ pub async fn manage(rt: Arc<CmfRuntime>, cfg: CmfCfg) {
         }
     }
 
-    // 3. Spawn the local server.
+    // 3. Spawn each server (fast).
+    for server in &servers {
+        if !std::path::Path::new(&server.model).exists() {
+            rt.fail_server(
+                &server.id,
+                format!("model file not found: {}", server.model),
+            );
+            continue;
+        }
         rt.log(format!(
             "starting: cortiq serve {} --host {} --port {}",
-        cfg.local_model, cfg.local_host, cfg.local_port
+            server.model, cfg.local_host, server.port
         ));
-    match spawn_serve(&cfg) {
+        match spawn_serve(server, &cfg) {
-        Ok(child) => rt.set_child(child),
+            Ok(child) => rt.add_child(server.id.clone(), child),
-        Err(e) => {
+            Err(e) => rt.fail_server(&server.id, format!("failed to spawn: {e}")),
-            rt.fail(format!("failed to spawn cortiq serve: {e}"));
-            return;
         }
     }
 
-    // 4. Wait until it is serving (model load can take a while for big models).
+    // 4. Wait until each spawned server is serving (model load can be slow).
-    if wait_healthy(&cfg.local_host, cfg.local_port, 240).await {
+    for server in &servers {
-        rt.status.lock().unwrap().healthy = true;
+        if !rt.is_running(&server.id) {
+            continue; // failed to spawn / file missing
+        }
+        if wait_healthy(&cfg.local_host, server.port, 240).await {
+            rt.set_healthy(&server.id, true);
             rt.log(format!(
-            "local CMF server ready — registered as model '{}' at http://{}:{}/v1",
+                "local CMF server ready — '{}' at http://{}:{}/v1",
-            cfg.model_id, cfg.local_host, cfg.local_port
+                server.id, cfg.local_host, server.port
             ));
         } else {
-        rt.fail("local CMF server did not become healthy in time");
+            rt.fail_server(&server.id, "did not become healthy in time");
+        }
     }
 }

src/config.rs

@@ -67,6 +67,44 @@ pub struct Config {
     pub cmf: CmfCfg,
 }
 
+/// One managed local model: the gateway runs a dedicated `cortiq serve` for it
+/// on its own port and registers it in the pool under `id`.
+#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
+pub struct CmfServer {
+    /// Pool id it is registered under (e.g. "cmf-qwen3-5-2b").
+    pub id: String,
+    /// Path to the `.cmf` to serve.
+    pub model: String,
+    /// TCP port the dedicated `cortiq serve` binds to (must be unique).
+    pub port: u16,
+    /// Worker threads (CMF_THREADS). 0 = runtime default.
+    #[serde(default = "default_cmf_threads")]
+    pub threads: usize,
+    /// Offload matvecs to the GPU (CMF_GPU=1).
+    #[serde(default)]
+    pub gpu: bool,
+    /// Serve-time O(1) attention override (exported as CMF_O1). Empty = follow
+    /// the hint recorded in the .cmf; "off" = force exact attention (fixes a
+    /// too-aggressive conversion without re-converting); "all"/"deepN"/"i,j,k"
+    /// = force-enable on those layers.
+    #[serde(default)]
+    pub o1: String,
+    /// Default generation params for this model — applied by the gateway when
+    /// the incoming request does not set them (an explicit request value wins).
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub temperature: Option<f32>,
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub top_p: Option<f32>,
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub max_tokens: Option<u32>,
+    /// Reasoning switch for thinking models (Qwen3/3.5 …): "off" injects
+    /// enable_thinking=false into requests that don't set it themselves —
+    /// the model answers directly, skipping the <think> phase. Needs
+    /// cortiq ≥ 0.2.1. Empty = model/template default.
+    #[serde(default)]
+    pub thinking: String,
+}
+
 /// CMF-format model factory: import a HuggingFace model, convert it to a
 /// local `.cmf` (quantized), and serve it via cortiq-server. Points the
 /// gateway at the Python converter + runtime on the host.
@@ -129,6 +167,40 @@ pub struct CmfCfg {
     /// (helps prefill of long prompts). Off by default.
     #[serde(default)]
     pub gpu: bool,
+    /// Managed local models. When non-empty (and `manage_server` is on), the
+    /// gateway spawns and supervises one `cortiq serve` per entry — each on its
+    /// own port — and registers each in the pool. Empty = fall back to the single
+    /// legacy `local_model` / `local_port` / `model_id` below (back-compat).
+    #[serde(default)]
+    pub servers: Vec<CmfServer>,
+}
+impl CmfCfg {
+    /// The managed local models actually in effect: the `servers` list when set,
+    /// otherwise the single legacy `local_model` (as a one-element list). Empty
+    /// when `manage_server` is off or nothing is configured.
+    pub fn effective_servers(&self) -> Vec<CmfServer> {
+        if !self.manage_server {
+            return Vec::new();
+        }
+        if !self.servers.is_empty() {
+            return self.servers.clone();
+        }
+        if !self.local_model.trim().is_empty() {
+            return vec![CmfServer {
+                id: self.model_id.clone(),
+                model: self.local_model.clone(),
+                port: self.local_port,
+                threads: self.threads,
+                gpu: self.gpu,
+                o1: String::new(),
+                temperature: None,
+                top_p: None,
+                max_tokens: None,
+                thinking: String::new(),
+            }];
+        }
+        Vec::new()
+    }
 }
 impl Default for CmfCfg {
     fn default() -> Self {
@@ -149,6 +221,7 @@ impl Default for CmfCfg {
             local_only: false,
             threads: default_cmf_threads(),
             gpu: false,
+            servers: Vec::new(),
         }
     }
 }
@@ -605,7 +678,8 @@ impl Config {
         // A managed local CMF model counts as an available model even though it
         // is registered dynamically rather than via a [[models]] entry — this
         // is what lets the gateway run on local models alone (no cloud, no router).
-        let has_local = self.cmf.manage_server && !self.cmf.local_model.trim().is_empty();
+        let servers = self.cmf.effective_servers();
+        let has_local = !servers.is_empty();
         if self.models.is_empty() && !has_local {
             // Fresh install: start with an EMPTY model pool so the admin console
             // comes up — the user adds models in /admin → Models (or points [cmf]
@@ -624,8 +698,25 @@ impl Config {
         // every target in [routing] must exist in the model pool
         let mut ids: std::collections::HashSet<&str> =
             self.models.iter().map(|m| m.id.as_str()).collect();
-        if has_local {
+        // managed local servers are known ids too; also enforce unique ports/ids
-            ids.insert(self.cmf.model_id.as_str());
+        let mut ports = std::collections::HashSet::new();
+        let mut managed_seen = std::collections::HashSet::new();
+        for s in &servers {
+            if !managed_seen.insert(s.id.as_str()) {
+                anyhow::bail!(
+                    "config: two managed local models share the id '{}' — give one of the rows a different id",
+                    s.id
+                );
+            }
+            if !ids.insert(s.id.as_str()) {
+                anyhow::bail!(
+                    "config: managed local model '{}' has the same id as a pool model — rename the row or delete the pool entry",
+                    s.id
+                );
+            }
+            if !ports.insert(s.port) {
+                anyhow::bail!("config: two managed models share port {}", s.port);
+            }
         }
         if !self.routing.default.is_empty() && !ids.contains(self.routing.default.as_str()) {
             anyhow::bail!(

src/import.rs

@@ -71,6 +71,20 @@ pub struct ImportParams {
     pub shard_max_gb: Option<f32>,
     #[serde(default)]
     pub skip_mtp: bool,
+    /// O(1) Nyström attention hint (native converter, cortiq ≥ 0.2.0):
+    /// `all` | `deepN` | `i,j,k`. Weights pass through unchanged — the runtime
+    /// reads the hint at load. Empty/None/"off" = exact attention.
+    #[serde(default)]
+    pub o1: Option<String>,
+    /// Landmark budget for the o1 hint (validated default 32).
+    #[serde(default)]
+    pub o1_m: Option<usize>,
+    /// Exact-window width for the o1 hint (validated default 128).
+    #[serde(default)]
+    pub o1_window: Option<usize>,
+    /// Permanent exact sink keys for the o1 hint (validated default 4).
+    #[serde(default)]
+    pub o1_sink: Option<usize>,
 }
 fn default_quant() -> String {
     "Q8_2F".into()
@@ -269,6 +283,34 @@ pub fn start_import(store: Arc<JobStore>, cfg: &CmfCfg, p: ImportParams) -> Resu
             cfg.converter
         ));
     }
+    // O(1) attention hint — native-converter feature (cortiq ≥ 0.2.0). Fail fast
+    // with a clear message instead of a mid-job "unexpected argument" error.
+    let o1 =
+        p.o1.as_deref()
+            .map(str::trim)
+            .filter(|s| !s.is_empty() && *s != "off")
+            .map(str::to_string);
+    if o1.is_some() {
+        if use_python {
+            return Err(
+                "O(1) attention is supported by the native converter only — remove the \
+                 linear-core / v-bit shape / mean-bits options to use it"
+                    .into(),
+            );
+        }
+        if p.repo.to_ascii_lowercase().contains("gguf") {
+            return Err("O(1) attention is not supported for GGUF imports yet".into());
+        }
+        let ver = crate::cmf_runtime::installed_version(&cfg.cortiq_bin);
+        if let Some(v) = &ver {
+            if crate::cmf_runtime::version_lt(v, "0.2.0") {
+                return Err(format!(
+                    "O(1) attention needs cortiq ≥ 0.2.0 (installed: {v}) — update the runtime \
+                     in Settings → Local models"
+                ));
+            }
+        }
+    }
     let _ = std::fs::create_dir_all(&cfg.models_dir);
     let base = if p.name.trim().is_empty() {
         sanitize(p.repo.rsplit('/').next().unwrap_or(&p.repo))
@@ -384,6 +426,23 @@ pub fn start_import(store: Arc<JobStore>, cfg: &CmfCfg, p: ImportParams) -> Resu
             a.push("--hf-token".into());
             a.push(t.clone());
         }
+        // O(1) attention hint (weights unchanged; the runtime reads it at load)
+        if let Some(spec) = &o1 {
+            a.push("--o1".into());
+            a.push(spec.clone());
+            if let Some(m) = p.o1_m {
+                a.push("--o1-m".into());
+                a.push(m.to_string());
+            }
+            if let Some(w) = p.o1_window {
+                a.push("--o1-window".into());
+                a.push(w.to_string());
+            }
+            if let Some(s) = p.o1_sink {
+                a.push("--o1-sink".into());
+                a.push(s.to_string());
+            }
+        }
         (
             cfg.cortiq_bin.clone(),
             a,

src/pipeline.rs

@@ -301,6 +301,35 @@ impl Pipeline {
         Ok((decision, route_source, final_candidates))
     }
 
+    /// Fill in the model's default generation params (from its Local-models row)
+    /// for anything the request left unset — an explicit request value wins.
+    fn apply_gen_defaults(req: &mut ChatRequest, cfg: &crate::config::Config, model_id: &str) {
+        if let Some(s) = cfg
+            .cmf
+            .effective_servers()
+            .into_iter()
+            .find(|s| s.id == model_id)
+        {
+            if req.params.temperature.is_none() {
+                req.params.temperature = s.temperature;
+            }
+            if req.params.top_p.is_none() {
+                req.params.top_p = s.top_p;
+            }
+            if req.params.max_tokens.is_none() {
+                req.params.max_tokens = s.max_tokens;
+            }
+            // reasoning switch: "off" → enable_thinking=false (cortiq ≥ 0.2.1
+            // renders the chat template with an empty <think> block); a client
+            // that sets enable_thinking itself wins.
+            if s.thinking == "off" && !req.params.passthrough.contains_key("enable_thinking") {
+                req.params
+                    .passthrough
+                    .insert("enable_thinking".into(), serde_json::Value::Bool(false));
+            }
+        }
+    }
+
     async fn run_inner(&self, req: ChatRequest, state: &SharedState) -> Result<ChatResponse> {
         let live = state.live();
         let (decision, route_source, mut final_candidates) = self.resolve(&req, &live).await?;
@@ -341,7 +370,9 @@ impl Pipeline {
                 route_source
             );
 
-            match provider.chat(req.clone()).await {
+            let mut attempt_req = req.clone();
+            Self::apply_gen_defaults(&mut attempt_req, &live.cfg, &model_id);
+            match provider.chat(attempt_req).await {
                 Ok(mut resp) => {
                     let cost =
                         provider.price(resp.usage.prompt_tokens, resp.usage.completion_tokens);
@@ -415,7 +446,9 @@ impl Pipeline {
                 model_id,
                 route_source
             );
-            match provider.chat_stream(req.clone()).await {
+            let mut attempt_req = req.clone();
+            Self::apply_gen_defaults(&mut attempt_req, &live.cfg, &model_id);
+            match provider.chat_stream(attempt_req).await {
                 Ok(stream) => {
                     let info = RouteInfo {
                         task_label: decision.task_label.clone(),

src/providers/openai.rs

@@ -45,6 +45,9 @@ impl OpenAiProvider {
         if let Some(m) = req.params.max_tokens {
             body["max_tokens"] = m.into();
         }
+        if let Some(p) = req.params.top_p {
+            body["top_p"] = p.into();
+        }
         if !req.tools.is_empty() {
             body["tools"] = serde_json::Value::Array(req.tools.clone());
         }

src/registry.rs

@@ -21,24 +21,27 @@ impl Registry {
                 .as_ref()
                 .and_then(|name| secrets.resolve(name));
             let provider: Arc<dyn Provider> = match m.provider.as_str() {
-                "openai" | "ollama" | "http" => Arc::new(OpenAiProvider::new(m, api_key)?),
+                // openrouter + lmstudio are OpenAI-compatible (Bearer + /chat/completions)
+                "openai" | "ollama" | "http" | "openrouter" | "lmstudio" => {
+                    Arc::new(OpenAiProvider::new(m, api_key)?)
+                }
                 "anthropic" => Arc::new(AnthropicProvider::new(m, api_key)?),
                 other => anyhow::bail!("unknown provider '{}' for model '{}'", other, m.id),
             };
             by_id.insert(m.id.clone(), provider);
         }
 
-        // Managed local CMF server: register it as an OpenAI-compatible provider
+        // Managed local CMF servers: register each as an OpenAI-compatible
-        // so it can be routed to like any other backend — including when the
+        // provider so it can be routed to like any other backend — including when
-        // router is unavailable or the gateway is in local-only mode.
+        // the router is unavailable or the gateway is in local-only mode.
-        if cfg.cmf.manage_server
+        for s in cfg.cmf.effective_servers() {
-            && !cfg.cmf.local_model.trim().is_empty()
+            if by_id.contains_key(&s.id) {
-            && !by_id.contains_key(&cfg.cmf.model_id)
+                continue; // a static [[models]] entry with the same id wins
-        {
+            }
             let m = ModelCfg {
-                id: cfg.cmf.model_id.clone(),
+                id: s.id.clone(),
                 provider: "openai".into(),
-                base_url: format!("http://{}:{}/v1", cfg.cmf.local_host, cfg.cmf.local_port),
+                base_url: format!("http://{}:{}/v1", cfg.cmf.local_host, s.port),
                 model: "cortiq".into(),
                 cost_tier: "local".into(),
                 price_in: 0.0,

src/state.rs

@@ -49,8 +49,8 @@ impl Live {
     /// (local-only / disabled) or unavailable, so CMF models serve with no router.
     pub fn local_candidates(&self) -> Vec<String> {
         let mut v = Vec::new();
-        if self.cfg.cmf.manage_server && !self.cfg.cmf.local_model.trim().is_empty() {
+        for s in self.cfg.cmf.effective_servers() {
-            v.push(self.cfg.cmf.model_id.clone());
+            v.push(s.id);
         }
         let def = self.routing.default_model().to_string();
         if !def.is_empty() && !v.contains(&def) {

src/stats.rs

@@ -83,9 +83,23 @@ impl Totals {
         self.cost_usd += r.cost_usd;
         self.latency_ms_sum += r.latency_ms;
     }
+    /// Accumulate another Totals into this one (for windowed aggregation).
+    fn merge(&mut self, o: &Totals) {
+        self.requests += o.requests;
+        self.ok += o.ok;
+        self.errors += o.errors;
+        self.failovers += o.failovers;
+        self.prompt_tokens += o.prompt_tokens;
+        self.completion_tokens += o.completion_tokens;
+        self.cost_usd += o.cost_usd;
+        self.latency_ms_sum += o.latency_ms_sum;
+    }
 }
 
-/// Per-minute bucket for time-series charts.
+/// Per-minute bucket. Beyond the chart fields it carries the per-group breakdown
+/// and failovers so the admin snapshot can be computed for ANY time window by
+/// summing the in-window buckets (the internal fields are not serialized into
+/// the `series` payload).
 #[derive(Clone, Debug, Default, Serialize)]
 pub struct Bucket {
     pub minute: u64, // unix seconds, rounded down to the minute
@@ -95,15 +109,25 @@ pub struct Bucket {
     pub completion_tokens: u64,
     pub cost_usd: f64,
     pub latency_ms_sum: u64,
+    #[serde(skip)]
+    pub failovers: u64,
+    #[serde(skip)]
+    pub by_model: HashMap<String, Totals>,
+    #[serde(skip)]
+    pub by_account: HashMap<String, Totals>,
+    #[serde(skip)]
+    pub by_tier: HashMap<String, Totals>,
+    #[serde(skip)]
+    pub by_label: HashMap<String, Totals>,
 }
 
 #[derive(Default)]
 struct Inner {
+    /// All-time totals + per-model (used by the Prometheus counters, which must
+    /// be monotonic — the windowed admin snapshot is derived from the buckets).
     total: Totals,
     by_model: HashMap<String, Totals>,
-    by_account: HashMap<String, Totals>,
+    /// Per-minute buckets — the single source for the windowed admin snapshot.
-    by_tier: HashMap<String, Totals>,
-    by_label: HashMap<String, Totals>,
     buckets: VecDeque<Bucket>,
     recent: VecDeque<RequestRecord>,
 }
@@ -166,31 +190,14 @@ impl Stats {
     }
 
     fn apply_aggregates(inner: &mut Inner, rec: &RequestRecord, retention_secs: u64) {
+        // all-time counters for Prometheus; the windowed snapshot is derived
+        // from the buckets below.
         inner.total.apply(rec);
         inner
             .by_model
             .entry(rec.model_id.clone())
             .or_default()
             .apply(rec);
-        inner
-            .by_account
-            .entry(if rec.account.is_empty() {
-                "anonymous".to_string()
-            } else {
-                rec.account.clone()
-            })
-            .or_default()
-            .apply(rec);
-        inner
-            .by_tier
-            .entry(rec.tier.clone())
-            .or_default()
-            .apply(rec);
-        inner
-            .by_label
-            .entry(rec.task_label.clone())
-            .or_default()
-            .apply(rec);
 
         let minute = rec.ts - (rec.ts % 60);
         match inner.buckets.back_mut() {
@@ -250,13 +257,31 @@ impl Stats {
             .filter(|b| b.minute >= cutoff)
             .collect();
 
-        let breakdown_src = match groupby {
+        // Windowed totals + breakdown = sum of the in-window buckets, so the KPIs
-            "account" => &inner.by_account,
+        // and the breakdown recompute whenever the range changes.
-            "tier" => &inner.by_tier,
+        let mut win = Totals::default();
-            "label" => &inner.by_label,
+        let mut group: HashMap<String, Totals> = HashMap::new();
-            _ => &inner.by_model,
+        for b in &series {
+            win.requests += b.requests;
+            win.errors += b.errors;
+            win.ok += b.requests.saturating_sub(b.errors);
+            win.failovers += b.failovers;
+            win.prompt_tokens += b.prompt_tokens;
+            win.completion_tokens += b.completion_tokens;
+            win.cost_usd += b.cost_usd;
+            win.latency_ms_sum += b.latency_ms_sum;
+            let src = match groupby {
+                "account" => &b.by_account,
+                "tier" => &b.by_tier,
+                "label" => &b.by_label,
+                _ => &b.by_model,
             };
-        let mut breakdown: Vec<serde_json::Value> = breakdown_src
+            for (k, t) in src {
+                group.entry(k.clone()).or_default().merge(t);
+            }
+        }
+
+        let mut breakdown: Vec<serde_json::Value> = group
             .iter()
             .map(|(k, t)| {
                 serde_json::json!({
@@ -280,17 +305,17 @@ impl Stats {
 
         serde_json::json!({
             "totals": {
-                "requests": inner.total.requests,
+                "requests": win.requests,
-                "ok": inner.total.ok,
+                "ok": win.ok,
-                "errors": inner.total.errors,
+                "errors": win.errors,
-                "failovers": inner.total.failovers,
+                "failovers": win.failovers,
-                "prompt_tokens": inner.total.prompt_tokens,
+                "prompt_tokens": win.prompt_tokens,
-                "completion_tokens": inner.total.completion_tokens,
+                "completion_tokens": win.completion_tokens,
-                "total_tokens": inner.total.prompt_tokens + inner.total.completion_tokens,
+                "total_tokens": win.prompt_tokens + win.completion_tokens,
-                "cost_usd": inner.total.cost_usd,
+                "cost_usd": win.cost_usd,
-                "avg_latency_ms": avg(inner.total.latency_ms_sum, inner.total.requests),
+                "avg_latency_ms": avg(win.latency_ms_sum, win.requests),
-                "success_rate": if inner.total.requests > 0 {
+                "success_rate": if win.requests > 0 {
-                    inner.total.ok as f64 / inner.total.requests as f64
+                    win.ok as f64 / win.requests as f64
                 } else { 0.0 },
             },
             "groupby": groupby,
@@ -299,6 +324,18 @@ impl Stats {
         })
     }
 
+    /// Reset all in-memory stats and truncate the JSONL log — the "clear logs"
+    /// admin action.
+    pub fn clear(&self) {
+        {
+            let mut inner = self.inner.lock().unwrap();
+            *inner = Inner::default();
+        }
+        if let Some(path) = &self.file {
+            let _ = std::fs::write(path, "");
+        }
+    }
+
     /// Recent requests (newest first), with pagination.
     pub fn recent(&self, limit: usize, offset: usize) -> Vec<RequestRecord> {
         let inner = self.inner.lock().unwrap();
@@ -358,10 +395,25 @@ fn fill_bucket(b: &mut Bucket, r: &RequestRecord) {
     if r.outcome != "ok" {
         b.errors += 1;
     }
+    if r.failover {
+        b.failovers += 1;
+    }
     b.prompt_tokens += r.prompt_tokens as u64;
     b.completion_tokens += r.completion_tokens as u64;
     b.cost_usd += r.cost_usd;
     b.latency_ms_sum += r.latency_ms;
+    // per-group breakdown, so the snapshot can be windowed by summing buckets
+    b.by_model.entry(r.model_id.clone()).or_default().apply(r);
+    b.by_account
+        .entry(if r.account.is_empty() {
+            "anonymous".to_string()
+        } else {
+            r.account.clone()
+        })
+        .or_default()
+        .apply(r);
+    b.by_tier.entry(r.tier.clone()).or_default().apply(r);
+    b.by_label.entry(r.task_label.clone()).or_default().apply(r);
 }
 
 fn avg(sum: u64, n: u64) -> u64 {

web/js/api.js

@@ -60,7 +60,8 @@ export const api = {
   listModels: () => j("GET", "/models"),
   createModel: (b) => j("POST", "/models", b),
   updateModel: (id, b) => j("PUT", "/models/" + encodeURIComponent(id), b),
-  deleteModel: (id) => j("DELETE", "/models/" + encodeURIComponent(id)),
+  deleteModel: (id, deleteFile = true) =>
+    j("DELETE", "/models/" + encodeURIComponent(id) + "?delete_file=" + (deleteFile ? "true" : "false")),
   probeModel: (id) => j("POST", "/models/" + encodeURIComponent(id) + "/probe"),
 
   // routing
@@ -87,6 +88,7 @@ export const api = {
 
   // stats
   stats: (range, groupby) => j("GET", `/stats?range=${range}&groupby=${groupby}`),
+  clearStats: () => j("DELETE", "/stats"),
   requests: (limit = 50, offset = 0) => j("GET", `/requests?limit=${limit}&offset=${offset}`),
 
   // playground
@@ -99,6 +101,8 @@ export const api = {
   cmfStatus: () => j("GET", "/cmf"),
   cmfInstall: () => j("POST", "/cmf/install"),
   cmfPortCheck: (port) => j("GET", "/cmf/port?port=" + encodeURIComponent(port)),
+  cmfFiles: () => j("GET", "/cmf/files"),
+  providerModels: (b) => j("POST", "/provider/models", b),
 
   // CMF model factory (HuggingFace → local .cmf)
   hfSearch: (q, limit = 24) =>

web/js/i18n/en.js

@@ -59,6 +59,9 @@ export default {
   // dashboard
   "dash.title": "Dashboard",
   "dash.subtitle": "Live traffic, cost and routing analytics for your model pool.",
+  "dash.clear": "Clear logs",
+  "dash.clearConfirm": "Clear all recorded stats and logs? This cannot be undone.",
+  "dash.cleared": "Logs cleared",
   "dash.range": "Range",
   "dash.range.1h": "1h",
   "dash.range.24h": "24h",
@@ -129,6 +132,20 @@ export default {
   "models.form.secretMissing": "not set",
   "models.form.secretPlaceholder": "set or replace key (write-only)",
   "models.deleteConfirm": "Delete model \"{id}\"?",
+  "models.deleteFile": "also delete the model file from disk",
+  "models.deletedWithFile": "Deleted (file removed from disk)",
+  "models.form.idOrModel": "Enter an id or a model name",
+  "models.form.apiKey": "API Key",
+  "models.form.apiKeyPlaceholder": "paste your API key",
+  "models.form.apiKeyNotNeeded": "not required (local server)",
+  "models.form.apiKeyHint": "Stored securely on the gateway; it is not shown again.",
+  "models.form.keyStored": "A key is saved — leave blank to keep it, or paste a new one.",
+  "models.form.keyFromEnv": "Using a key from an environment variable.",
+  "models.form.testKey": "Test & load models",
+  "models.form.testing": "checking…",
+  "models.form.foundModels": "{n} models loaded",
+  "models.form.modelPlaceholder": "type to search…",
+  "models.form.modelHint": "Click “Test & load models” to pull the list from your key, then pick one.",
   "models.managed": "managed",
   "models.managedHint": "Provided by the managed CMF server (Settings → Local models). Configure it there.",
 
@@ -206,6 +223,12 @@ export default {
   "pg.model": "Model",
   "pg.temp": "Temperature",
   "pg.maxTokens": "Max tokens",
+  "pg.think": "Thoughts",
+  "pg.think.auto": "auto (model default)",
+  "pg.think.off": "off — answer directly",
+  "pg.think.on": "on — force reasoning",
+  "pg.thinkHelp": "Reasoning phase of thinking models. Auto = model/row default; off/on override for this request (needs cortiq >= 0.2.1 for local models).",
+  "pg.topPHelp": "Nucleus sampling: only tokens within this cumulative probability are sampled. Empty = default.",
   "pg.send": "Send",
   "pg.stream": "Stream",
   "pg.sending": "Sending…",
@@ -244,9 +267,23 @@ export default {
   "settings.cmf.threads": "Worker threads",
   "settings.cmf.gpu": "Use GPU (Metal / wgpu)",
   "settings.cmf.speedNote": "Threads parallelise the decode matvecs (~3x faster; 0 = runtime default). GPU offload mainly helps prefill of long prompts — decode is memory-bandwidth-bound, so on Apple Silicon it is roughly neutral.",
+  "settings.cmf.pick": "select a model",
+  "settings.cmf.models": "Local models",
+  "settings.cmf.modelsHelp": "Each row is one model the gateway runs (its own cortiq serve on its own port) and registers in the pool. Add as many as you have .cmf files and free ports.",
+  "settings.cmf.rowHint": "id · .cmf file · port · threads · GPU · O(1) · T°/top-p/max tok. Applied on save: a removed row stops its server (memory freed), a new row starts one.",
+  "settings.cmf.thinkAuto": "thoughts: auto",
+  "settings.cmf.thinkOff": "thoughts: off",
+  "settings.cmf.thinkHelp": "Reasoning models (Qwen3/3.5) write out a thinking phase before the answer. Off = answer directly without it (faster, shorter; needs cortiq >= 0.2.1). A request that sets enable_thinking itself always wins.",
+  "settings.cmf.tempHelp": "Default temperature for this model (0 = deterministic, 0.7 = balanced, >1 = creative). Used when the request doesn't set one; a request value always wins. Empty = provider default.",
+  "settings.cmf.topPHelp": "Default nucleus sampling (top-p) for this model: only tokens within this cumulative probability are sampled (0.9 typical). Empty = provider default.",
+  "settings.cmf.maxTokHelp": "Default max output tokens for this model when the request doesn't set one. Protects against runaway long generations. Empty = provider default.",
+  "settings.cmf.o1Auto": "O(1): auto",
+  "settings.cmf.o1Help": "O(1) attention at serve time: auto — follow the hint recorded in the .cmf; off — force exact attention (fixes incoherent output after a too-aggressive --o1 conversion, no re-convert needed); all/deepN — force-enable on those layers.",
+  "settings.cmf.addServer": "+ Add local model",
   "settings.cmf.checkPort": "Check port",
   "settings.cmf.portFree": "port is free",
   "settings.cmf.portBusy": "port is in use",
+  "settings.cmf.portOwn": "in use by this gateway's current server — freed on restart, replacing is fine",
   "settings.cmf.useSuggested": "use {port}",
   "settings.cmf.serving": "serving {model} on :{port}",
   "settings.cmf.intro": "Run models locally with the cortiq engine — works even offline or when the router is down.",
@@ -299,6 +336,10 @@ export default {
   "settings.route.maxChars": "Max chars",
   "settings.route.cacheTtl": "Cache TTL",
   "settings.route.profile": "Default profile",
+  "settings.route.strategyHelp": "What text is sent to the router for classification: last_user = only the last user message; last_user_plus_system = last user message + system prompt; concat_all = the whole conversation.",
+  "settings.route.maxCharsHelp": "Max characters of that text sent to the router (longer is truncated). More = more accurate classification, but slower/pricier router calls.",
+  "settings.route.cacheTtlHelp": "How long a routing decision is cached for an identical request before the router is asked again (e.g. 60s, 5m).",
+  "settings.route.profileHelp": "Default routing profile when a request doesn't specify one: cost-saver (cheapest), balanced, or quality-first (best model). Biases which model the router picks.",
   "settings.breaker": "Circuit breaker",
   "settings.breaker.threshold": "Error threshold",
   "settings.breaker.cooldown": "Cooldown",
@@ -376,4 +417,21 @@ export default {
   "import.adv.shard": "Shard max (GB)",
   "import.adv.noShard": "no sharding",
   "import.adv.skipMtp": "skip MTP head",
+  "import.adv.o1": "O(1) attention",
+  "import.adv.o1Hint": "Constant-memory streaming attention: the KV cache is released after prefill, so long contexts don't grow memory. Weights stay byte-identical — only a load-time hint is recorded; you can switch back anytime (serve with --o1 off). Validated on Qwen3-0.6B: all layers ≈ ×1.18 perplexity.",
+  "import.adv.o1Off": "off — exact attention (default)",
+  "import.adv.o1All": "all — convert every softmax layer",
+  "import.adv.o1Deep": "deep N — only the N deepest layers",
+  "import.adv.o1Custom": "custom — explicit layer list",
+  "import.adv.o1DeepN": "N deep layers",
+  "import.adv.o1Layers": "Layers (comma-separated)",
+  "import.adv.o1mHint": "landmark budget (default 32)",
+  "import.adv.o1mLabel": "Landmarks (--o1-m)",
+  "import.adv.o1mHelp": "How many landmark tokens summarize the whole older history when attention is approximated. More = closer to exact attention but slower and more memory; fewer = faster but rougher. 32 is the validated sweet spot — change only if you know why.",
+  "import.adv.o1windowLabel": "Exact window (--o1-window)",
+  "import.adv.o1windowHelp": "The last N tokens always get exact (non-approximated) attention. This protects quality on the text the model just wrote/read. Bigger = better local quality, more memory. Default 128.",
+  "import.adv.o1sinkLabel": "Sink keys (--o1-sink)",
+  "import.adv.o1sinkHelp": "The first N tokens of the context are kept exact forever ('attention sinks'). Models lean on these anchors — keeping them exact stabilizes long generations. Default 4; rarely needs changing.",
+  "import.adv.o1windowHint": "exact-window width (default 128)",
+  "import.adv.o1sinkHint": "permanent exact sink keys (default 4)",
 };

web/js/i18n/ru.js

@@ -58,6 +58,9 @@ export default {
   // dashboard
   "dash.title": "Дашборд",
   "dash.subtitle": "Трафик в реальном времени, расходы и аналитика роутинга по пулу моделей.",
+  "dash.clear": "Очистить логи",
+  "dash.clearConfirm": "Очистить всю статистику и логи? Отменить нельзя.",
+  "dash.cleared": "Логи очищены",
   "dash.range": "Период",
   "dash.range.1h": "1ч",
   "dash.range.24h": "24ч",
@@ -128,6 +131,20 @@ export default {
   "models.form.secretMissing": "не задан",
   "models.form.secretPlaceholder": "задать или заменить ключ (только запись)",
   "models.deleteConfirm": "Удалить модель \"{id}\"?",
+  "models.deleteFile": "также удалить файл модели с диска",
+  "models.deletedWithFile": "Удалено (файл стёрт с диска)",
+  "models.form.idOrModel": "Укажите id или имя модели",
+  "models.form.apiKey": "API-ключ",
+  "models.form.apiKeyPlaceholder": "вставьте API-ключ",
+  "models.form.apiKeyNotNeeded": "не нужен (локальный сервер)",
+  "models.form.apiKeyHint": "Хранится безопасно на шлюзе; больше не показывается.",
+  "models.form.keyStored": "Ключ сохранён — оставьте пустым, чтобы не менять, или вставьте новый.",
+  "models.form.keyFromEnv": "Используется ключ из переменной окружения.",
+  "models.form.testKey": "Проверить и загрузить модели",
+  "models.form.testing": "проверка…",
+  "models.form.foundModels": "загружено моделей: {n}",
+  "models.form.modelPlaceholder": "введите для поиска…",
+  "models.form.modelHint": "Нажмите «Проверить и загрузить модели», чтобы подтянуть список по ключу, затем выберите.",
   "models.managed": "управляемая",
   "models.managedHint": "Предоставляется CMF-сервером (Настройки → Локальные модели). Настраивается там.",
 
@@ -205,6 +222,12 @@ export default {
   "pg.model": "Модель",
   "pg.temp": "Температура",
   "pg.maxTokens": "Макс. токены",
+  "pg.think": "Мысли",
+  "pg.think.auto": "авто (по настройке модели)",
+  "pg.think.off": "выкл — отвечать сразу",
+  "pg.think.on": "вкл — принудительно размышлять",
+  "pg.thinkHelp": "Фаза размышлений reasoning-моделей. Авто = настройка модели; выкл/вкл переопределяют для этого запроса (для локальных нужен cortiq ≥ 0.2.1).",
+  "pg.topPHelp": "Nucleus-сэмплинг: выбираются токены в пределах этой суммарной вероятности. Пусто = по умолчанию.",
   "pg.send": "Отправить",
   "pg.stream": "Стриминг",
   "pg.sending": "Отправка…",
@@ -243,9 +266,23 @@ export default {
   "settings.cmf.threads": "Рабочие потоки",
   "settings.cmf.gpu": "Использовать GPU (Metal / wgpu)",
   "settings.cmf.speedNote": "Потоки распараллеливают декод-матвеки (~×3; 0 = по умолчанию рантайма). GPU-оффлоад помогает в основном на prefill длинных промптов — декод упирается в пропускную способность памяти, поэтому на Apple Silicon почти нейтрален.",
+  "settings.cmf.pick": "выберите модель",
+  "settings.cmf.models": "Локальные модели",
+  "settings.cmf.modelsHelp": "Каждая строка — отдельная модель: шлюз поднимает свой cortiq serve на своём порту и регистрирует её в пуле. Добавляйте столько, сколько есть .cmf-файлов и свободных портов.",
+  "settings.cmf.rowHint": "id · файл .cmf · порт · потоки · GPU · O(1) · T°/top-p/max tok. Применяется сразу при сохранении: удалённая строка останавливает сервер (память освобождается), новая — запускает.",
+  "settings.cmf.thinkAuto": "мысли: авто",
+  "settings.cmf.thinkOff": "мысли: выкл",
+  "settings.cmf.thinkHelp": "Рассуждающие модели (Qwen3/3.5) сначала пишут фазу размышлений. Выкл = отвечать сразу, без неё (быстрее и короче; нужен cortiq ≥ 0.2.1). Если запрос сам задал enable_thinking — он важнее.",
+  "settings.cmf.tempHelp": "Температура по умолчанию для этой модели (0 = детерминированно, 0.7 = баланс, >1 = креативно). Подставляется, если запрос её не задал; значение из запроса всегда важнее. Пусто = дефолт провайдера.",
+  "settings.cmf.topPHelp": "Nucleus-сэмплинг (top-p) по умолчанию: выбираются только токены в пределах этой суммарной вероятности (обычно 0.9). Пусто = дефолт провайдера.",
+  "settings.cmf.maxTokHelp": "Максимум выходных токенов по умолчанию, если запрос не задал свой. Защищает от бесконечных генераций. Пусто = дефолт провайдера.",
+  "settings.cmf.o1Auto": "O(1): авто",
+  "settings.cmf.o1Help": "O(1)-внимание при запуске: авто — использовать подсказку из .cmf; off — принудительно точное внимание (лечит бессвязный вывод после слишком агрессивной --o1 конвертации, без переконвертации); all/deepN — принудительно включить на этих слоях.",
+  "settings.cmf.addServer": "+ Добавить модель",
   "settings.cmf.checkPort": "Проверить порт",
   "settings.cmf.portFree": "порт свободен",
   "settings.cmf.portBusy": "порт занят",
+  "settings.cmf.portOwn": "занят текущим сервером шлюза — освободится после перезапуска, замена допустима",
   "settings.cmf.useSuggested": "занять {port}",
   "settings.cmf.serving": "работает: {model} на :{port}",
   "settings.cmf.intro": "Запускайте модели локально движком cortiq — работает даже офлайн или когда роутер недоступен.",
@@ -298,6 +335,10 @@ export default {
   "settings.route.maxChars": "Макс. символов",
   "settings.route.cacheTtl": "TTL кэша",
   "settings.route.profile": "Профиль по умолчанию",
+  "settings.route.strategyHelp": "Какой текст отправляется роутеру для классификации: last_user — только последнее сообщение пользователя; last_user_plus_system — последнее сообщение + системный промпт; concat_all — вся переписка.",
+  "settings.route.maxCharsHelp": "Максимум символов этого текста, отправляемого роутеру (длиннее — обрезается). Больше = точнее классификация, но медленнее/дороже вызовы роутера.",
+  "settings.route.cacheTtlHelp": "Сколько кэшируется решение роутинга для одинакового запроса, прежде чем роутер спрашивается снова (напр. 60s, 5m).",
+  "settings.route.profileHelp": "Профиль роутинга по умолчанию, если запрос его не задаёт: cost-saver (дешевле), balanced (баланс) или quality-first (качество). Влияет на выбор модели роутером.",
   "settings.breaker": "Автоматический выключатель",
   "settings.breaker.threshold": "Порог ошибок",
   "settings.breaker.cooldown": "Время восстановления",
@@ -375,4 +416,21 @@ export default {
   "import.adv.shard": "Шард макс (ГБ)",
   "import.adv.noShard": "без шардинга",
   "import.adv.skipMtp": "пропустить MTP-голову",
+  "import.adv.o1": "O(1)-внимание",
+  "import.adv.o1Hint": "Потоковое внимание с постоянной памятью: KV-кэш освобождается после prefill, длинный контекст не раздувает память. Веса не меняются — записывается лишь подсказка на загрузку; вернуться можно всегда (запуск с --o1 off). Проверено на Qwen3-0.6B: все слои ≈ ×1.18 перплексии.",
+  "import.adv.o1Off": "выкл — точное внимание (по умолчанию)",
+  "import.adv.o1All": "all — конвертировать все softmax-слои",
+  "import.adv.o1Deep": "deep N — только N самых глубоких слоёв",
+  "import.adv.o1Custom": "вручную — явный список слоёв",
+  "import.adv.o1DeepN": "N глубоких слоёв",
+  "import.adv.o1Layers": "Слои (через запятую)",
+  "import.adv.o1mHint": "число опорных точек (по умолчанию 32)",
+  "import.adv.o1mLabel": "Опорные точки (--o1-m)",
+  "import.adv.o1mHelp": "Сколько опорных токенов представляют всю старую историю при аппроксимации внимания. Больше — ближе к точному вниманию, но медленнее и больше памяти; меньше — быстрее, но грубее. 32 — проверенное значение; меняйте, только если понимаете зачем.",
+  "import.adv.o1windowLabel": "Точное окно (--o1-window)",
+  "import.adv.o1windowHelp": "Последние N токенов всегда обрабатываются точным (не аппроксимированным) вниманием — это защищает качество на только что написанном/прочитанном тексте. Больше — лучше локальное качество, больше памяти. По умолчанию 128.",
+  "import.adv.o1sinkLabel": "Sink-ключи (--o1-sink)",
+  "import.adv.o1sinkHelp": "Первые N токенов контекста навсегда остаются точными («якоря внимания»). Модели опираются на эти якоря — их точность стабилизирует длинную генерацию. По умолчанию 4; менять почти никогда не нужно.",
+  "import.adv.o1windowHint": "ширина точного окна (по умолчанию 128)",
+  "import.adv.o1sinkHint": "постоянные точные sink-ключи (по умолчанию 4)",
 };

web/js/ui.js

@@ -25,7 +25,9 @@ export function clear(node) {
 
 export function mount(node, ...children) {
   clear(node);
-  for (const c of children.flat()) if (c != null && c !== false) node.appendChild(c);
+  for (const c of children.flat())
+    if (c != null && c !== false)
+      node.appendChild(typeof c === "object" ? c : document.createTextNode(String(c)));
   return node;
 }
 

web/js/views/dashboard.js

@@ -1,5 +1,5 @@
 // Dashboard — KPIs, time-series charts, breakdown, recent requests, health.
-import { h, mount, money, num, ms, pct, timeAgo } from "../ui.js";
+import { h, mount, money, num, ms, pct, timeAgo, toast, confirmDialog } from "../ui.js";
 import { t } from "../i18n.js";
 import { api } from "../api.js";
 import { lineChart, barList } from "../charts.js";
@@ -53,10 +53,24 @@ export async function renderDashboard() {
           "div",
           { class: "flex" },
           h("div", { class: "grow" }, h("h2", {}, t("dash.title")), h("p", {}, t("dash.subtitle"))),
+          h("div", { class: "flex", style: "gap:8px;align-items:center" },
             seg(RANGES, range, (r) => {
               range = r;
               load();
-          }, (r) => t("dash.range." + r))
+            }, (r) => t("dash.range." + r)),
+            h("button", {
+              class: "btn sm danger",
+              onClick: async () => {
+                if (!(await confirmDialog(t("dash.clearConfirm")))) return;
+                try {
+                  await api.clearStats();
+                  toast(t("dash.cleared"), "good");
+                  await load();
+                } catch (e) {
+                  toast(String(e.message || e), "bad");
+                }
+              },
+            }, t("dash.clear")))
         )
       ),
 

web/js/views/import.js

@@ -15,8 +15,14 @@ const QUANTS = [
   ["F32", "import.q.f32"],
 ];
 
-function field(label, control, hint) {
+// A small "?" badge with a hover/focus tooltip (same style as Settings).
-  return h("label", { class: "field" }, h("span", {}, label), control,
+function help(text) {
+  return text ? h("span", { class: "help", title: text, tabindex: "0", role: "img", "aria-label": text }, "?") : null;
+}
+function field(label, control, hint, helpText) {
+  return h("label", { class: "field" },
+    h("span", {}, label, helpText ? " " : null, help(helpText)),
+    control,
     hint ? h("div", { class: "hint" }, hint) : null);
 }
 function opt(v, label, sel) {
@@ -125,6 +131,32 @@ export function renderImport() {
     const meanBitsIn = h("input", { type: "number", step: "0.1", placeholder: "5.0" });
     const shardIn = h("input", { type: "number", step: "0.5", placeholder: t("import.adv.noShard") });
     const skipMtp = h("input", { type: "checkbox" });
+    // O(1) attention (cortiq ≥ 0.2.0): constant-memory streaming attention hint,
+    // weights byte-identical — the runtime applies it at load.
+    const o1Sel = h("select", {},
+      opt("", t("import.adv.o1Off")),
+      opt("all", t("import.adv.o1All")),
+      opt("deep", t("import.adv.o1Deep")),
+      opt("custom", t("import.adv.o1Custom")));
+    const o1DeepIn = h("input", { type: "number", min: "1", placeholder: "12" });
+    const o1CustomIn = h("input", { placeholder: "0,4,8" });
+    const o1M = h("input", { type: "number", min: "1", placeholder: "32" });
+    const o1Window = h("input", { type: "number", min: "0", placeholder: "128" });
+    const o1Sink = h("input", { type: "number", min: "0", placeholder: "4" });
+    const o1SpecRow = h("div", { class: "row", style: "display:none" },
+      field(t("import.adv.o1DeepN"), o1DeepIn),
+      field(t("import.adv.o1Layers"), o1CustomIn));
+    const o1Knobs = h("div", { class: "row", style: "display:none" },
+      field(t("import.adv.o1mLabel"), o1M, t("import.adv.o1mHint"), t("import.adv.o1mHelp")),
+      field(t("import.adv.o1windowLabel"), o1Window, t("import.adv.o1windowHint"), t("import.adv.o1windowHelp")),
+      field(t("import.adv.o1sinkLabel"), o1Sink, t("import.adv.o1sinkHint"), t("import.adv.o1sinkHelp")));
+    o1Sel.addEventListener("change", () => {
+      const on = !!o1Sel.value;
+      o1Knobs.style.display = on ? "" : "none";
+      o1SpecRow.style.display = o1Sel.value === "deep" || o1Sel.value === "custom" ? "" : "none";
+      o1DeepIn.parentElement.style.display = o1Sel.value === "deep" ? "" : "none";
+      o1CustomIn.parentElement.style.display = o1Sel.value === "custom" ? "" : "none";
+    });
     const advBody = h(
       "div",
       { class: "adv-body", style: "display:none" },
@@ -137,6 +169,9 @@ export function renderImport() {
       h("div", { class: "row" },
         field(t("import.adv.shard"), shardIn),
         field("MTP", h("label", { class: "check" }, skipMtp, t("import.adv.skipMtp")))),
+      field(t("import.adv.o1"), o1Sel, t("import.adv.o1Hint")),
+      o1SpecRow,
+      o1Knobs,
     );
     const advToggle = h("button", { class: "btn ghost sm" }, "⚙ " + t("import.adv.title"));
     advToggle.addEventListener("click", () => {
@@ -156,6 +191,19 @@ export function renderImport() {
       if (vbitSel.value) params.vbit_shape = vbitSel.value;
       if (meanBitsIn.value) params.mean_bits = parseFloat(meanBitsIn.value);
       if (shardIn.value) params.shard_max_gb = parseFloat(shardIn.value);
+      if (o1Sel.value) {
+        // spec: all | deepN | explicit layer list
+        params.o1 =
+          o1Sel.value === "deep" ? "deep" + (parseInt(o1DeepIn.value) || 12)
+          : o1Sel.value === "custom" ? o1CustomIn.value.trim()
+          : "all";
+        if (!params.o1) delete params.o1;
+        else {
+          if (o1M.value) params.o1_m = parseInt(o1M.value);
+          if (o1Window.value) params.o1_window = parseInt(o1Window.value);
+          if (o1Sink.value) params.o1_sink = parseInt(o1Sink.value);
+        }
+      }
       goBtn.disabled = true;
       try {
         await api.startImport(params);

web/js/views/models.js

@@ -1,5 +1,5 @@
 // Models — pool table with add/edit modal, probe, and per-model secret entry.
-import { h, mount, modal, toast, confirmDialog, money } from "../ui.js";
+import { h, mount, modal, toast, money } from "../ui.js";
 import { t } from "../i18n.js";
 import { api } from "../api.js";
 import { appState } from "../app.js";
@@ -16,37 +16,126 @@ function keyBadge(src) {
   return h("span", { class: "badge " + (map[src] ?? "") }, src);
 }
 
+// Verified per-provider defaults (base URL, key env, whether a key is needed).
+const PROVIDER_DEFAULTS = {
+  openai:     { base: "https://api.openai.com/v1",    keyEnv: "OPENAI_API_KEY",     needsKey: true,  tier: "mid",       caps: ["tools", "vision"] },
+  anthropic:  { base: "https://api.anthropic.com",    keyEnv: "ANTHROPIC_API_KEY",  needsKey: true,  tier: "expensive", caps: ["tools", "vision"] },
+  openrouter: { base: "https://openrouter.ai/api/v1", keyEnv: "OPENROUTER_API_KEY", needsKey: true,  tier: "mid",       caps: ["tools", "vision"] },
+  lmstudio:   { base: "http://localhost:1234/v1",     keyEnv: "",                   needsKey: false, tier: "local",     caps: ["tools", "vision"] },
+  ollama:     { base: "http://localhost:11434/v1",    keyEnv: "",                   needsKey: false, tier: "local",     caps: ["tools"] },
+  http:       { base: "",                              keyEnv: "",                   needsKey: false, tier: "cheap",     caps: ["tools"] },
+};
+
+// Derive a clean model id from a model name (last path segment, slugified).
+// e.g. "qwen/qwen3.7-max" → "qwen3-7-max".
+function slugId(name) {
+  const base = (name || "").split("/").pop() || "";
+  return base.toLowerCase().replace(/[^a-z0-9]+/g, "-").replace(/^-+|-+$/g, "");
+}
+
 function modelForm(meta, existing) {
-  const m = existing || { provider: "openai", kind: "chat", cost_tier: "cheap", caps: [], price_in: 0, price_out: 0 };
+  const m = existing || { provider: "openai", kind: "chat", cost_tier: "mid", caps: ["tools"], price_in: 0, price_out: 0 };
-  const idIn = h("input", { value: m.id || "", placeholder: "local-qwen", disabled: existing ? true : null });
+  const idIn = h("input", { value: m.id || "", placeholder: "my-model", disabled: existing ? true : null });
   const providerSel = h("select", {}, ...(meta.providers || ["openai"]).map((p) => opt(p, p, p === m.provider)));
-  const baseUrlIn = h("input", { value: m.base_url || "", placeholder: "http://localhost:8000/v1" });
+  const baseUrlIn = h("input", { value: m.base_url || "", placeholder: "https://api.openai.com/v1" });
-  const modelIn = h("input", { value: m.model || "", placeholder: "qwen2.5-7b-instruct" });
+
+  // API key — typed directly, stored securely as a secret under a provider env name.
+  const keyIn = h("input", { type: "password", placeholder: t("models.form.apiKeyPlaceholder"), autocomplete: "new-password" });
+  const testBtn = h("button", { class: "btn small", type: "button" }, t("models.form.testKey"));
+  const keyStatus = h("span", { class: "small", style: "margin-left:8px" });
+
+  // Model — below the key (some providers list models from the key); searchable datalist.
+  const dlId = "model-opts-" + (m.id || "new");
+  const dl = h("datalist", { id: dlId });
+  if (m.model) dl.appendChild(h("option", { value: m.model }));
+  const modelIn = h("input", { value: m.model || "", placeholder: t("models.form.modelPlaceholder"), list: dlId, autocomplete: "off" });
+
   const kindSel = h("select", {}, ...(meta.kinds || ["chat"]).map((k) => opt(k, k, k === m.kind)));
   const tierSel = h("select", {}, ...(meta.cost_tiers || ["cheap"]).map((c) => opt(c, c, c === m.cost_tier)));
   const priceInIn = h("input", { type: "number", step: "0.01", value: m.price_in ?? 0 });
   const priceOutIn = h("input", { type: "number", step: "0.01", value: m.price_out ?? 0 });
-  const keyEnvIn = h("input", { value: m.api_key_env || "", placeholder: "OPENAI_API_KEY" });
-  const secretIn = h("input", { type: "password", placeholder: t("models.form.secretPlaceholder") });
   const capInputs = (meta.caps || ["tools", "vision"]).map((c) => {
     const cb = h("input", { type: "checkbox", checked: (m.caps || []).includes(c) ? true : null });
     return { c, cb };
   });
 
+  const envFor = () =>
+    (existing && existing.api_key_env) ||
+    (PROVIDER_DEFAULTS[providerSel.value] || {}).keyEnv ||
+    providerSel.value.toUpperCase().replace(/[^A-Z0-9]/g, "_") + "_API_KEY";
+
+  // Fill in a provider's verified defaults when it is selected.
+  function applyDefaults() {
+    const d = PROVIDER_DEFAULTS[providerSel.value];
+    if (!d) return;
+    baseUrlIn.value = d.base;
+    tierSel.value = d.tier;
+    capInputs.forEach(({ c, cb }) => { cb.checked = d.caps.includes(c); });
+    keyIn.placeholder = d.needsKey ? t("models.form.apiKeyPlaceholder") : t("models.form.apiKeyNotNeeded");
+    mount(keyStatus, "");
+  }
+  providerSel.addEventListener("change", applyDefaults);
+  if (!existing) applyDefaults();
+
+  // Auto-fill the id from the model name (until the user edits id by hand).
+  if (!existing) {
+    modelIn.addEventListener("input", () => {
+      if (!idIn.value || idIn._auto) {
+        idIn.value = slugId(modelIn.value);
+        idIn._auto = true;
+      }
+    });
+    idIn.addEventListener("input", () => { idIn._auto = false; });
+  }
+
+  async function testKey() {
+    if (!baseUrlIn.value.trim()) { mount(keyStatus, h("span", { style: "color:var(--bad)" }, t("models.form.baseUrl") + "?")); return; }
+    testBtn.disabled = true;
+    mount(keyStatus, t("models.form.testing"));
+    try {
+      const r = await api.providerModels({
+        provider: providerSel.value,
+        base_url: baseUrlIn.value.trim(),
+        api_key: keyIn.value.trim() || null,
+        api_key_env: envFor(),
+      });
+      if (r.ok) {
+        dl.replaceChildren(...(r.models || []).map((mm) => h("option", { value: mm })));
+        mount(keyStatus, h("span", { style: "color:var(--good)" }, "✓ " + t("models.form.foundModels", { n: r.count })));
+        modelIn.focus();
+      } else {
+        mount(keyStatus, h("span", { style: "color:var(--bad)" }, "✗ " + (r.error || t("common.error"))));
+      }
+    } catch (e) {
+      mount(keyStatus, h("span", { style: "color:var(--bad)" }, "✗ " + String(e.message || e)));
+    }
+    testBtn.disabled = false;
+  }
+  testBtn.addEventListener("click", testKey);
+
+  const keyHint =
+    existing && existing.key_source === "store" ? t("models.form.keyStored")
+    : existing && existing.key_source === "env" ? t("models.form.keyFromEnv")
+    : t("models.form.apiKeyHint");
+
   const node = h(
     "div",
     {},
     h("div", { class: "row" }, field(t("models.form.id"), idIn, existing ? null : t("models.form.idHint")), field(t("models.form.provider"), providerSel)),
     field(t("models.form.baseUrl"), baseUrlIn),
-    field(t("models.form.model"), modelIn),
+    field(t("models.form.apiKey"), h("div", { class: "flex wrap", style: "gap:8px;align-items:center" }, keyIn, testBtn, keyStatus), keyHint),
+    field(t("models.form.model"), modelIn, t("models.form.modelHint")),
+    dl,
     h("div", { class: "row" }, field(t("models.form.kind"), kindSel), field(t("models.form.costTier"), tierSel)),
     h("div", { class: "row" }, field(t("models.form.priceIn"), priceInIn), field(t("models.form.priceOut"), priceOutIn)),
-    field(t("models.form.caps"), h("div", { class: "flex wrap" }, ...capInputs.map(({ c, cb }) => h("label", { class: "check" }, cb, c)))),
+    field(t("models.form.caps"), h("div", { class: "flex wrap" }, ...capInputs.map(({ c, cb }) => h("label", { class: "check" }, cb, c))))
-    field(t("models.form.apiKeyEnv"), keyEnvIn, t("models.form.apiKeyEnvHint")),
-    field(t("models.form.secret"), secretIn)
   );
 
-  const getValue = () => ({
+  const getValue = () => {
+    const typedKey = keyIn.value.trim();
+    const hasKey = !!typedKey || !!(existing && existing.api_key_env);
+    const keyEnv = envFor();
+    return {
       model: {
         id: existing ? m.id : idIn.value.trim(),
         provider: providerSel.value,
@@ -56,12 +145,13 @@ function modelForm(meta, existing) {
         cost_tier: tierSel.value,
         price_in: parseFloat(priceInIn.value) || 0,
         price_out: parseFloat(priceOutIn.value) || 0,
-      api_key_env: keyEnvIn.value.trim() || null,
+        api_key_env: hasKey ? keyEnv : null,
         caps: capInputs.filter(({ cb }) => cb.checked).map(({ c }) => c),
       },
-    secret: secretIn.value,
+      secret: typedKey,
-    keyEnv: keyEnvIn.value.trim(),
+      keyEnv,
-  });
+    };
+  };
 
   return { node, getValue };
 }
@@ -70,12 +160,14 @@ async function openModal(meta, existing, reload) {
   const { node, getValue } = modelForm(meta, existing);
   modal(existing ? t("models.form.editTitle") : t("models.form.addTitle"), node, async () => {
     const { model, secret, keyEnv } = getValue();
+    // Forgot the id? Derive it from the model name so saving never errors on a
+    // blank id.
+    if (!model.id) model.id = slugId(model.model);
     if (!model.id) {
-      toast(t("models.form.id") + "?", "bad");
+      toast(t("models.form.idOrModel"), "bad");
       return false;
     }
-    // strip null api_key_env so serde keeps it optional
+    if (!model.api_key_env) delete model.api_key_env; // keep serde field optional
-    if (!model.api_key_env) delete model.api_key_env;
     try {
       if (existing) await api.updateModel(existing.id, model);
       else await api.createModel(model);
@@ -188,15 +280,33 @@ function row(m, meta, reload) {
           "button",
           {
             class: "btn sm danger",
-            onClick: async () => {
+            onClick: () => {
-              if (!(await confirmDialog(t("models.deleteConfirm", { id: m.id })))) return;
+              // local models may have a .cmf on disk — let the user decide its fate
+              const isLocal = m.managed || /127\.0\.0\.1|localhost/.test(m.base_url || "");
+              // unchecked by default: deleting a pool entry should not silently
+              // cost the user a multi-GB re-download/conversion
+              const fileCb = h("input", { type: "checkbox" });
+              const body = h(
+                "div",
+                {},
+                h("p", {}, t("models.deleteConfirm", { id: m.id })),
+                isLocal
+                  ? h("label", { class: "check" }, fileCb, t("models.deleteFile"))
+                  : null
+              );
+              modal(t("common.delete") + ": " + m.id, body, async () => {
                 try {
-                await api.deleteModel(m.id);
+                  const r = await api.deleteModel(m.id, isLocal && fileCb.checked);
-                toast(t("toast.deleted"), "good");
+                  toast(
-                reload();
+                    r.file_removed ? t("models.deletedWithFile") : t("toast.deleted"),
+                    "good"
+                  );
+                  await reload();
                 } catch (e) {
                   toast(String(e.message || e), "bad");
+                  return false;
                 }
+              }, t("common.delete"));
             },
           },
           t("common.delete")

web/js/views/playground.js

@@ -16,6 +16,12 @@ export async function renderPlayground() {
   const promptIn = h("textarea", { placeholder: t("pg.promptPlaceholder"), rows: 5 });
   const tempIn = h("input", { type: "number", step: "0.1", value: 0.7 });
   const maxTokIn = h("input", { type: "number", value: 512 });
+  const topPIn = h("input", { type: "number", step: "0.05", min: 0, max: 1, placeholder: "0.9", title: t("pg.topPHelp") });
+  // reasoning switch: auto = model/row default; off/on override per request
+  const thinkSel = h("select", { title: t("pg.thinkHelp") },
+    h("option", { value: "" }, t("pg.think.auto")),
+    h("option", { value: "off" }, t("pg.think.off")),
+    h("option", { value: "on" }, t("pg.think.on")));
   const streamCb = h("input", { type: "checkbox", checked: true });
   const profileSel = h("select", {}, ...(meta.profiles || []).map((p) => h("option", { value: p }, p)));
   const modelSel = h("select", {}, ...ids.map((id) => h("option", { value: id }, id)));
@@ -69,6 +75,8 @@ export async function renderPlayground() {
       temperature: parseFloat(tempIn.value),
       max_tokens: parseInt(maxTokIn.value) || null,
     };
+    if (topPIn.value !== "") body.top_p = parseFloat(topPIn.value);
+    if (thinkSel.value) body.enable_thinking = thinkSel.value === "on";
     try {
       if (streamCb.checked) {
         const started = performance.now();
@@ -151,6 +159,9 @@ export async function renderPlayground() {
           h("label", { class: "field" }, h("span", {}, t("pg.mode")), modeSeg),
           modeExtra,
           h("div", { class: "row" }, h("label", { class: "field" }, h("span", {}, t("pg.temp")), tempIn), h("label", { class: "field" }, h("span", {}, t("pg.maxTokens")), maxTokIn)),
+          h("div", { class: "row" },
+            h("label", { class: "field" }, h("span", {}, "top-p"), topPIn),
+            h("label", { class: "field" }, h("span", {}, t("pg.think")), thinkSel)),
           h(
             "div",
             { class: "flex", style: "justify-content:space-between" },

web/js/views/settings.js

@@ -207,11 +207,87 @@ export async function renderSettings() {
   const cmfManage = check(t("settings.cmf.manage"), s.cmf?.manage_server, t("settings.cmf.manageHelp"));
   const cmfAutoInstall = check(t("settings.cmf.autoInstall"), s.cmf?.auto_install, t("settings.cmf.autoInstallHelp"));
   const cmfAutoUpdate = check(t("settings.cmf.autoUpdate"), s.cmf?.auto_update, t("settings.cmf.autoUpdateHelp"));
-  const cmfModel = h("input", { value: s.cmf?.local_model || "", placeholder: "models/my-model.cmf" });
+  // ── managed local models: one supervised `cortiq serve` per row ──
-  const cmfPort = h("input", { type: "number", value: s.cmf?.local_port ?? 8081 });
+  const serverRows = [];
-  const cmfThreads = h("input", { type: "number", min: 0, value: s.cmf?.threads ?? 8 });
+  const serversWrap = h("div", {});
-  const cmfGpu = check(t("settings.cmf.gpu"), s.cmf?.gpu, t("settings.cmf.gpuHelp"));
+  const legacyServers = s.cmf?.local_model
-  const cmfModelIdVal = s.cmf?.model_id || "cmf-local"; // auto — no manual "model id" field
+    ? [{ id: s.cmf.model_id || "cmf-local", model: s.cmf.local_model, port: s.cmf.local_port ?? 8081, threads: s.cmf.threads ?? 8, gpu: !!s.cmf.gpu }]
+    : [];
+  const initServers = (s.cmf?.servers && s.cmf.servers.length) ? s.cmf.servers : legacyServers;
+  // available .cmf files under models_dir → offer as a dropdown, not a typed path
+  const cmfFiles = ((await api.cmfFiles().catch(() => ({ files: [] }))).files) || [];
+  const fileLabel = (f) => f.name + (f.size ? ` (${(f.size / 1073741824).toFixed(f.size >= 1073741824 ? 1 : 2)} GB)` : f.is_dir ? " (dir)" : "");
+  function autoId(path) {
+    const base = (path || "").split("/").pop().replace(/\.cmf$/i, "");
+    return base ? "cmf-" + base.replace(/[^a-zA-Z0-9_-]/g, "-") : "";
+  }
+  function makeServerRow(sv) {
+    sv = sv || { id: "", model: "", port: 8090, threads: 8, gpu: false };
+    const idIn = h("input", { value: sv.id || "", placeholder: "cmf-id", style: "width:130px" });
+    // pick from existing .cmf files rather than typing a path
+    const modelSel = h("select", { style: "flex:2;min-width:150px" });
+    const seen = new Set();
+    modelSel.appendChild(h("option", { value: "" }, "— " + t("settings.cmf.pick") + " —"));
+    cmfFiles.forEach((f) => { modelSel.appendChild(h("option", { value: f.path }, fileLabel(f))); seen.add(f.path); });
+    if (sv.model && !seen.has(sv.model)) modelSel.appendChild(h("option", { value: sv.model }, sv.model));
+    modelSel.value = sv.model || "";
+    const portIn = h("input", { type: "number", value: sv.port ?? 8090, style: "width:82px" });
+    const threadsIn = h("input", { type: "number", min: 0, value: sv.threads ?? 8, style: "width:60px", title: t("settings.cmf.threadsHelp") });
+    const gpuIn = h("input", { type: "checkbox", checked: sv.gpu ? true : null, title: t("settings.cmf.gpuHelp") });
+    // serve-time O(1) override: off cures a too-aggressive --o1 conversion
+    const o1Opts = ["", "off", "all", "deep4", "deep8", "deep12"];
+    const o1In = h("select", { style: "width:110px", title: t("settings.cmf.o1Help") },
+      ...o1Opts.map((v) => h("option", { value: v, selected: (sv.o1 || "") === v ? true : null },
+        v === "" ? t("settings.cmf.o1Auto") : v)));
+    if (sv.o1 && !o1Opts.includes(sv.o1)) o1In.appendChild(h("option", { value: sv.o1, selected: true }, sv.o1));
+    // reasoning switch for thinking models (off → enable_thinking=false)
+    const thinkIn = h("select", { style: "width:118px", title: t("settings.cmf.thinkHelp") },
+      h("option", { value: "", selected: !sv.thinking ? true : null }, t("settings.cmf.thinkAuto")),
+      h("option", { value: "off", selected: sv.thinking === "off" ? true : null }, t("settings.cmf.thinkOff")));
+    // per-model generation defaults (applied when the request doesn't set them)
+    const tempIn = h("input", { type: "number", step: "0.1", min: 0, max: 2, value: sv.temperature ?? "", placeholder: "T° 0.7", style: "width:74px", title: t("settings.cmf.tempHelp") });
+    const topPIn = h("input", { type: "number", step: "0.05", min: 0, max: 1, value: sv.top_p ?? "", placeholder: "top-p", style: "width:70px", title: t("settings.cmf.topPHelp") });
+    const maxTokIn = h("input", { type: "number", min: 1, value: sv.max_tokens ?? "", placeholder: "max tok", style: "width:82px", title: t("settings.cmf.maxTokHelp") });
+    const checkBtn = h("button", { class: "btn small", type: "button" }, t("settings.cmf.checkPort"));
+    const removeBtn = h("button", { class: "btn small danger", type: "button", title: t("common.delete") }, "✕");
+    const st = h("span", { class: "small" });
+    modelSel.addEventListener("change", () => { if (!idIn.value || idIn._auto) { idIn.value = autoId(modelSel.value); idIn._auto = true; } });
+    idIn.addEventListener("input", () => { idIn._auto = false; });
+    checkBtn.addEventListener("click", async () => {
+      const p = parseInt(portIn.value, 10); if (!p) return;
+      checkBtn.disabled = true; mount(st, t("status.checking"));
+      try {
+        const r = await api.cmfPortCheck(p);
+        if (r.available) mount(st, h("span", { style: "color:var(--good)" }, "✓ " + t("settings.cmf.portFree")));
+        else if (r.managed) {
+          // held by this gateway's own managed server — replacing the model on
+          // this port is fine, it frees up on restart
+          mount(st, h("span", { style: "color:var(--accent)" }, "ℹ " + t("settings.cmf.portOwn")));
+        } else {
+          const kids = [h("span", { style: "color:var(--bad)" }, "✗ " + t("settings.cmf.portBusy"))];
+          if (r.suggested) { const b = h("button", { class: "btn small", type: "button" }, t("settings.cmf.useSuggested", { port: r.suggested })); b.addEventListener("click", () => { portIn.value = r.suggested; mount(st, ""); }); kids.push(b); }
+          mount(st, h("span", {}, ...kids));
+        }
+      } catch (e) { mount(st, String(e.message || e)); }
+      checkBtn.disabled = false;
+    });
+    const row = h("div", { class: "flex wrap", style: "gap:6px;align-items:center;margin:6px 0" },
+      idIn, modelSel, portIn, checkBtn, threadsIn, h("label", { class: "flex", style: "gap:4px" }, gpuIn, "GPU"), o1In, thinkIn, tempIn, topPIn, maxTokIn, removeBtn, st);
+    const entry = { node: row, read: () => {
+      const v = { id: idIn.value.trim() || autoId(modelSel.value), model: modelSel.value.trim(), port: parseInt(portIn.value) || 8090, threads: parseInt(threadsIn.value) || 0, gpu: gpuIn.checked, o1: o1In.value, thinking: thinkIn.value };
+      if (tempIn.value !== "") v.temperature = parseFloat(tempIn.value);
+      if (topPIn.value !== "") v.top_p = parseFloat(topPIn.value);
+      if (maxTokIn.value !== "") v.max_tokens = parseInt(maxTokIn.value);
+      return v;
+    } };
+    removeBtn.addEventListener("click", () => { const i = serverRows.indexOf(entry); if (i >= 0) serverRows.splice(i, 1); row.remove(); });
+    serverRows.push(entry);
+    serversWrap.appendChild(row);
+    return entry;
+  }
+  initServers.forEach((sv) => makeServerRow(sv));
+  const addServerBtn = h("button", { class: "btn small", type: "button" }, t("settings.cmf.addServer"));
+  addServerBtn.addEventListener("click", () => makeServerRow());
   // "Install cortiq" — the CLI that RUNS the .cmf format. Must be present before
   // a local model server can start; this is step 1 for a new user.
   const cmfStatusEl = h("span", { class: "muted" }, "…");
@@ -228,14 +304,17 @@ export async function renderSettings() {
           ? h("span", {}, h("span", { class: "badge ok" }, "✓"), " cortiq " + (st.version || ""))
           : h("span", {}, h("span", { class: "badge bad" }, "×"), " " + t("settings.cmf.notInstalled"))
       );
-      // reflect the managed server state under the "run locally" toggle
+      // per-server serving indicators
-      if (st.manage_server && st.running) {
+      const servers = st.servers || [];
-        mount(cmfServingEl, h("span", { style: "color:var(--good)" },
+      if (!servers.length) {
-          "▶ " + t("settings.cmf.serving", { model: st.model_id || "", port: st.local_port }) + (st.healthy ? "" : " …")));
-      } else if (st.last_error) {
-        mount(cmfServingEl, h("span", { style: "color:var(--bad)" }, "✗ " + st.last_error));
-      } else {
         mount(cmfServingEl, "");
+      } else {
+        mount(cmfServingEl, h("div", {}, ...servers.map((sv) => {
+          if (sv.last_error) return h("div", { style: "color:var(--bad)" }, "✗ " + sv.id + ": " + sv.last_error);
+          if (sv.running) return h("div", { style: "color:var(--good)" },
+            "▶ " + t("settings.cmf.serving", { model: sv.id, port: sv.port }) + (sv.healthy ? "" : " …"));
+          return h("div", { class: "muted" }, "• " + sv.id + " :" + sv.port);
+        })));
       }
     } catch (_) {
       mount(cmfStatusEl, "—");
@@ -251,7 +330,6 @@ export async function renderSettings() {
         await new Promise((r) => setTimeout(r, 3000));
         const st = await api.cmfStatus();
         if (st.installed) { toast(t("settings.cmf.installedOk"), "good"); break; }
-        if (st.last_error) { toast(st.last_error, "bad"); break; }
       }
     } catch (e) {
       toast(String(e.message || e), "bad");
@@ -261,36 +339,6 @@ export async function renderSettings() {
   });
   refreshCmfStatus();
 
-  // Port availability check — let the user confirm the chosen port is free
-  // before enabling the managed server (the default 8081 often clashes).
-  const cmfPortStatus = h("span", { class: "small", style: "margin-left:8px" });
-  const cmfPortCheckBtn = h("button", { class: "btn small", type: "button" }, t("settings.cmf.checkPort"));
-  async function checkPort() {
-    const p = parseInt(cmfPort.value, 10);
-    if (!p) { mount(cmfPortStatus, ""); return; }
-    cmfPortCheckBtn.disabled = true;
-    mount(cmfPortStatus, t("status.checking"));
-    try {
-      const r = await api.cmfPortCheck(p);
-      if (r.available) {
-        mount(cmfPortStatus, h("span", { style: "color:var(--good)" }, "✓ " + t("settings.cmf.portFree")));
-      } else {
-        const kids = [h("span", { style: "color:var(--bad)" }, "✗ " + t("settings.cmf.portBusy"))];
-        if (r.suggested) {
-          const useBtn = h("button", { class: "btn small", type: "button", style: "margin-left:8px" },
-            t("settings.cmf.useSuggested", { port: r.suggested }));
-          useBtn.addEventListener("click", () => { cmfPort.value = r.suggested; checkPort(); });
-          kids.push(useBtn);
-        }
-        mount(cmfPortStatus, h("span", {}, ...kids));
-      }
-    } catch (e) {
-      mount(cmfPortStatus, String(e.message || e));
-    }
-    cmfPortCheckBtn.disabled = false;
-  }
-  cmfPortCheckBtn.addEventListener("click", checkPort);
-
   const saveBtn = h("button", { class: "btn primary" }, t("common.save"));
   saveBtn.addEventListener("click", async () => {
     saveBtn.disabled = true;
@@ -335,11 +383,9 @@ export async function renderSettings() {
         manage_server: cmfManage.cb.checked,
         auto_install: cmfAutoInstall.cb.checked,
         auto_update: cmfAutoUpdate.cb.checked,
-        local_model: cmfModel.value.trim(),
+        // the managed-models list is the source of truth; clear the legacy single
-        local_port: parseInt(cmfPort.value) || 8081,
+        servers: serverRows.map((r) => r.read()).filter((sv) => sv.model && sv.id),
-        model_id: cmfModelIdVal,
+        local_model: "",
-        threads: parseInt(cmfThreads.value) || 0,
-        gpu: cmfGpu.cb.checked,
       },
     };
     // serde: drop null optionals
@@ -420,9 +466,11 @@ export async function renderSettings() {
         "div",
         { class: "card" },
         h("div", { class: "card-head" }, h("h3", {}, t("settings.route"))),
-        field(t("settings.route.strategy"), strategySel),
+        field(t("settings.route.strategy"), strategySel, null, t("settings.route.strategyHelp")),
-        h("div", { class: "row" }, field(t("settings.route.maxChars"), maxChars), field(t("settings.route.cacheTtl"), cacheTtl)),
+        h("div", { class: "row" },
-        field(t("settings.route.profile"), profileSel)
+          field(t("settings.route.maxChars"), maxChars, null, t("settings.route.maxCharsHelp")),
+          field(t("settings.route.cacheTtl"), cacheTtl, null, t("settings.route.cacheTtlHelp"))),
+        field(t("settings.route.profile"), profileSel, null, t("settings.route.profileHelp"))
       ),
       h(
         "div",
@@ -477,24 +525,21 @@ export async function renderSettings() {
         ),
         h("div", { class: "row" }, cmfAutoInstall.node, cmfAutoUpdate.node),
         h("div", { class: "divider" }),
-        // Step 2 — add a model. Models are imported in the Models/Import tab; here
+        // Step 2 — managed local models: one supervised cortiq serve per row, each
-        // you just point the local server at the .cmf file it produced.
+        // on its own port. Add as many as you have .cmf files + free ports.
-        h("div", { class: "field" }, h("span", {}, "2. " + t("settings.cmf.addModel"), " ", help(t("settings.cmf.addModelHelp"))),
+        h("div", { class: "field" },
+          h("span", {}, "2. " + t("settings.cmf.models"), " ", help(t("settings.cmf.modelsHelp"))),
+          h("div", { class: "hint" }, t("settings.cmf.rowHint")),
           h("div", { class: "hint" }, t("settings.cmf.importNote"),
             " ", h("a", { href: "#/import" }, t("settings.cmf.importLink") + " ↗"))),
-        field(t("settings.cmf.model"), cmfModel, null, t("settings.cmf.modelHelp")),
+        serversWrap,
-        h("div", { class: "divider" }),
+        addServerBtn,
-        // Step 3 — run it locally + performance.
-        cmfManage.node,
         cmfServingEl,
-        h("div", { class: "row" },
+        h("div", { class: "hint" }, t("settings.cmf.speedNote")),
-          field(t("settings.cmf.port"),
+        h("div", { class: "divider" }),
-            h("div", { class: "flex wrap" }, cmfPort, cmfPortCheckBtn, cmfPortStatus),
+        // Step 3 — how the gateway uses them.
-            null, t("settings.cmf.portHelp")),
+        cmfManage.node,
-          field(t("settings.cmf.threads"), cmfThreads, null, t("settings.cmf.threadsHelp"))),
+        cmfLocalOnly.node
-        cmfGpu.node,
-        cmfLocalOnly.node,
-        h("div", { class: "hint" }, t("settings.cmf.speedNote"))
       )
     ),
     h(