feat(web): stream chat replies token-by-token (M3)

- llm.chat_call_stream: streaming generator for all 3 backends (Ollama NDJSON,
  OpenAI/MI50 SSE), accumulating tool-call fragments by index.
- chat.respond_stream: mirrors respond()'s tool loop and persistence/compaction,
  yielding ("delta", text) / ("tool", name) / ("done", reply).
- POST /v1/chat/stream: SSE endpoint; blocking generator bridged to async via a
  worker thread + asyncio.Queue. Old completions endpoint kept as fallback.
- Client streams into a live bubble with a blinking caret; rAF-throttled render
  (no full re-parse per token) and instant scroll during stream — fixes iOS
  Safari ghosting from per-token smooth-scroll. Falls back to the blocking
  endpoint only if nothing streamed (no double-persist).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

lyra/chat.py

@@ -162,3 +162,60 @@ def respond(session_id: str, user_msg: str, backend: Backend = "cloud",
     # Compact this session once enough new turns have piled up.
     summary.maybe_summarize(session_id)
     return reply
+
+
+def respond_stream(session_id: str, user_msg: str, backend: Backend = "cloud",
+                   model_override: str | None = None):
+    """Streaming generator version of `respond`.
+
+    Yields ("delta", text) as content streams in, and ("tool", name) when a tool
+    runs. Persists the full exchange and yields a final ("done", reply) — matching
+    `respond`'s side effects (memory + compaction) exactly.
+    """
+    cfg = config.load()
+    model = {"local": cfg.local_model, "cloud": cfg.chat_model, "mi50": cfg.mi50_model}.get(
+        backend, backend
+    )
+    if model_override and backend == "cloud":
+        model = model_override
+    logbus.log(
+        "info", "chat request (stream)", session=session_id, backend=backend,
+        model=model, embed=cfg.embed_backend,
+    )
+
+    messages = build_messages(session_id, user_msg)
+    tool_specs = toolkit.specs() if backend in TOOL_BACKENDS else None
+    ctx = {"session_id": session_id, "backend": backend}
+    parts: list[str] = []
+    for _ in range(MAX_TOOL_ROUNDS):
+        assistant_msg = None
+        tool_calls = None
+        for ev, payload in llm.chat_call_stream(
+            messages, backend=backend, model=model, tools=tool_specs
+        ):
+            if ev == "delta":
+                parts.append(payload)
+                yield ("delta", payload)
+            elif ev == "message":
+                assistant_msg = payload
+            elif ev == "tool_calls":
+                tool_calls = payload
+        if not tool_calls:
+            break
+        messages.append(assistant_msg)  # her tool-call request
+        for tc in tool_calls:
+            result = toolkit.dispatch(tc["name"], tc["arguments"], ctx)
+            logbus.log("info", "tool call", session=session_id, tool=tc["name"], result=result[:80])
+            messages.append({"role": "tool", "tool_call_id": tc["id"], "content": result})
+            yield ("tool", tc["name"])
+
+    reply = "".join(parts)
+    if not reply:
+        reply = "(I got tangled using my tools there — say that again?)"
+        yield ("delta", reply)
+    logbus.log("info", "reply", session=session_id, chars=len(reply))
+
+    memory.remember(session_id, "user", user_msg)
+    memory.remember(session_id, "assistant", reply)
+    summary.maybe_summarize(session_id)
+    yield ("done", reply)