turbollm 1.11.2 → 1.11.7

package/README.md

@@ -291,6 +291,13 @@ the detail:
   in chat (name, preferences, hardware) using your own loaded model, and carries them into future
   new conversations. Nothing leaves your device; the full fact list is reviewable and deletable
   from Settings → Memory, and turning the toggle off stops new chats from seeing it immediately.
+- **Turbo Link** *(experimental, off by default)* — link a second machine running TurboLLM
+  with one pasted link string and use its models as if they were local: they appear in the chat
+  model picker (grouped by machine), in `/v1/models` and in `launch --model`, and Models,
+  Downloads and Engines merge both machines' rows into one list. You grant an explicit capability
+  set per link — use models only, run a server, or full control — optionally pinned to specific
+  models, and revocable at any time from the host. Not yet verified against a real two-machine
+  setup, so it stays behind Settings → Experimental.
 - **Thinking-budget control** — a graduated slider, not just on/off: cap reasoning to a specific
   token count, disable it entirely, or leave it unlimited. Works in Chat and Code alike.
 - **Reasoning-effort control** for models whose own chat template supports it (e.g. Qwen3.8) —
@@ -355,7 +362,7 @@ there. Entirely local; nothing leaves your machine.
 </details>
 
 <details>
-<summary><strong>🔌 APIs &amp; integrations — OpenAI + Anthropic, plus a model-loading gateway</strong></summary>
+<summary><strong>🔌 APIs &amp; integrations — OpenAI + Anthropic, a model-loading gateway, and a stateful chat API</strong></summary>
 
 <br/>
 
@@ -375,6 +382,17 @@ curl http://127.0.0.1:6996/v1/chat/completions \
   Claude Code below. No other local host offers this.
 - **Structured output** — constrain any response to a **GBNF grammar** (or JSON shape).
 - **API-key auth** you can require when sharing over a LAN (Settings → Network & sharing).
+- **A stateful External Chat API** (**Experimental — manual setup, no onboarding UI yet**),
+  `/api/ext/v1`, for building a real chat product instead of the gateway's stateless
+  send-full-history-every-call model — persistence, history, message editing, and resumable
+  streaming generation are handled for you. Off by default; using it today needs two manual
+  `config.json` edits — a key with an explicit, non-`local` `tenant` field (the local tenant, this
+  install's own desktop UI data, is never reachable through this API), and a configured
+  `chatStore` adapter (the SQLite default only ever serves the `local` tenant). A generated
+  OpenAPI spec, a TypeScript client, and a working Postgres adapter example ship with it. The
+  daemon logs a warning at startup while it's enabled — a remote tenant's generations currently
+  inherit this install's own tool permissions, not an independent trust boundary of their own.
+  [Full guide](https://turbollm.dev/docs/api/external).
 
 **The gateway loads models for you.** Most local hosts make you load a model first, then call it.
 TurboLLM's gateway reads the `model` field of any incoming request, **fuzzy-matches it to your