Improve model status indicators and tune timeouts

Green/red emoji circles replace Unicode symbols for at-a-glance model
readiness. Increase connect timeouts (10→30s) for cold model loading,
warm timeout (120→300s) for large models, and pull timeout (600→1800s)
to match route-level limit. Show error toast on SSE reconnection failure
instead of silently recovering.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

frontend/src/components/ModelSelector.vue

@@ -49,7 +49,7 @@ onMounted(async () => {
       :key="m.name"
       :value="m.name"
     >
-      {{ isHot(m.name) ? "\u2B24 " : "\u25CB " }}{{ m.name }}
+      {{ isHot(m.name) ? "🟢 " : "🔴 " }}{{ m.name }}
     </option>
   </select>
 </template>

frontend/src/stores/chat.ts

@@ -239,6 +239,10 @@ export const useChatStore = defineStore("chat", () => {
 
     // Recovery fallback: re-fetch conversation from DB
     if (!gotDone && currentConversation.value?.id === convId) {
+      useToastStore().show(
+        "Connection lost — response may be incomplete",
+        "error",
+      );
       try {
         await fetchConversation(convId);
       } catch {

src/fabledassistant/routes/chat.py

@@ -269,7 +269,7 @@ async def warm_model_route():
 
     async def _warm():
         try:
-            async with httpx.AsyncClient(timeout=120.0) as client:
+            async with httpx.AsyncClient(timeout=300.0) as client:
                 await client.post(
                     f"{Config.OLLAMA_URL}/api/generate",
                     json={"model": model, "prompt": "", "keep_alive": "30m"},

src/fabledassistant/services/llm.py

@@ -54,7 +54,7 @@ async def ensure_model(model: str) -> None:
 
     logger.info("Pulling model '%s' from Ollama...", model)
     try:
-        async with httpx.AsyncClient(timeout=600.0) as client:
+        async with httpx.AsyncClient(timeout=1800.0) as client:
             async with client.stream(
                 "POST",
                 f"{Config.OLLAMA_URL}/api/pull",
@@ -80,7 +80,7 @@ async def stream_chat(
     payload: dict = {"model": model, "messages": messages, "stream": True}
     if options:
         payload["options"] = options
-    async with httpx.AsyncClient(timeout=httpx.Timeout(1800.0, connect=10.0, read=300.0)) as client:
+    async with httpx.AsyncClient(timeout=httpx.Timeout(1800.0, connect=30.0, read=300.0)) as client:
         async with client.stream(
             "POST",
             f"{Config.OLLAMA_URL}/api/chat",
@@ -100,7 +100,7 @@ async def stream_chat(
 
 async def generate_completion(messages: list[dict], model: str) -> str:
     """Non-streaming chat completion, returns full response text."""
-    async with httpx.AsyncClient(timeout=httpx.Timeout(1800.0, connect=10.0, read=300.0)) as client:
+    async with httpx.AsyncClient(timeout=httpx.Timeout(1800.0, connect=30.0, read=300.0)) as client:
         resp = await client.post(
             f"{Config.OLLAMA_URL}/api/chat",
             json={"model": model, "messages": messages, "stream": False},

summary.md

@@ -12,7 +12,7 @@
 >    Include file-level details in the commit body when the change is non-trivial.
 
 ## Last Updated
-2026-02-14 — Phase 6.0: Model selection, dashboard chat input, model warming
+2026-02-14 — Phase 6.1: Model status colors, timeout tuning, SSE error feedback
 
 ## Project Overview
 Fabled Assistant is a self-hosted note-taking and task-tracking application with
@@ -527,8 +527,10 @@ When adding a new migration, follow these conventions:
 - Per-conversation model selection via ModelSelector dropdown in chat header (persisted via PATCH)
 - Dashboard inline chat input: model selector + note picker + textarea; creates conversation and navigates
 - Model warming: default model pre-loaded into Ollama on dashboard mount via fire-and-forget POST
-- Hot/cold model indicators: `/api/chat/ps` proxies Ollama `/api/ps`; ModelSelector shows filled/empty circles
+- Hot/cold model indicators: `/api/chat/ps` proxies Ollama `/api/ps`; ModelSelector shows green/red emoji circles
 - Ollama configured with `OLLAMA_MAX_LOADED_MODELS=2` and `OLLAMA_KEEP_ALIVE=30m`
+- Timeout tuning: connect timeout 30s (cold model loading), warm timeout 300s, pull timeout 1800s
+- SSE reconnection failure shows error toast instead of silently recovering
 
 ### Authentication & User Management
 - Session cookie auth with bcrypt, first-user-is-admin, orphaned data claiming