turbollm 1.6.5 → 1.7.1

package/README.md

@@ -79,7 +79,7 @@ TurboLLM does the opposite:
   a **VRAM-fit verdict before you load** — no more flag guessing.
 - **📊 Real tokens/sec, never faked.** Speed in the model list is *measured on your machine*
   from actual generation — live while you chat, and remembered per model.
-- **🪶 Lightweight.** A ~0.3 MB npm package on Node — **no Electron, no bundled Chromium, no
+- **🪶 Lightweight.** A ~2 MB npm package on Node — **no Electron, no bundled Chromium, no
   Python**. It downloads only the engine your GPU actually needs (Vulkan ≈ 38 MB).
 - **🔌 Drop-in APIs.** OpenAI **and** Anthropic-compatible — so Claude Code and every existing
   tool work unchanged.
@@ -177,6 +177,8 @@ the detail:
 - **Saved per-model profiles, per engine** — tune once per (model, engine) pair, so
   switching engines (or between two installs of the same engine, e.g. a fork) never
   overwrites another engine's tuning for the same model.
+- **Configurable VRAM headroom** (Settings → Engine, 300 MB–2 GB, default 1 GB) — tell auto-tune
+  how much VRAM to keep free for other GPU workloads instead of a fixed margin.
 
 </details>
 
@@ -210,6 +212,9 @@ the detail:
   Atlassian, Neon, Supabase, Cloudflare, Zapier, Apify, Mixpanel) and open-source local MCPs
   (filesystem, git, postgres, playwright, …), plus your own custom servers. Connected tools appear
   in every chat with no restart. A **Research** persona forces multi-step web search and cites sources inline.
+- **Tool-call approval gate** — every tool call asks for your approval by default before it runs,
+  with **Deny**, **Allow**, **Allow for this chat**, or **Always Allow** on an inline bar above the
+  composer. Set per-tool defaults globally from Developer → Tool permissions.
 
 </details>
 
@@ -256,6 +261,12 @@ library, and loads it on the fly** if it isn't already running — then keeps up
 hot** in an LRU pool so the next switch is instant. An agent (or Claude Code) that hops between a
 coding model, a vision model, and an embedder just names each one and it works — no pre-wiring.
 
+**One command wires up more than just Claude Code.** `turbollm launch <cli>` also supports
+`opencode`, `kilo`, `openclaw`, and `hermes` — each gets pointed at TurboLLM the way that tool
+expects (config file merge, or its own CLI command) instead of a manual copy-paste setup. Inside
+Claude Code, `/model` lists your local models directly (with **Auto Model Swap** on in Settings →
+Gateway) so you can switch mid-session instead of only at launch.
+
 </details>
 
 <details>
@@ -284,7 +295,7 @@ connected); **Remove** undoes it.
 
 <br/>
 
-- A **~0.3 MB npm package** on Node — no Electron, no bundled Chromium, no Python.
+- A **~2 MB npm package** on Node — no Electron, no bundled Chromium, no Python.
 - **Offline-first** — no account, no backend, no internet, no telemetry.
 - **Windows · macOS · Linux**, with a CPU fallback when there's no GPU.
 
@@ -365,12 +376,20 @@ turbollm launch claude               # auto-loads a model if none is running, th
 turbollm launch claude --model qwen3-8b   # load a specific model first, then launch
 ```
 
-It sets Claude Code's `ANTHROPIC_BASE_URL` / `ANTHROPIC_MODEL` at TurboLLM and execs `claude`;
-extra args are forwarded. If no model is loaded it auto-loads your last-used one (or the first
-in your library); `--model` picks a specific one by key or name. If `claude` isn't installed,
-it tells you how. The in-app
-**Developer** screen also shows copy-paste env snippets for any OpenAI- or Anthropic-compatible
-tool (Open WebUI, Kilo Code, opencode, …).
+It sets Claude Code's `ANTHROPIC_BASE_URL` and pins `ANTHROPIC_MODEL` to whatever model is
+loaded (so the status line, `/status`, and context-window sizing all reflect the real model),
+then execs `claude`; extra args are forwarded. If no model is loaded it auto-loads your
+last-used one (or the first in your library); `--model` picks a specific one by key or name.
+If `claude` isn't installed, it tells you how.
+
+Other coding CLIs work too — `turbollm launch opencode`, `turbollm launch kilo`, and
+`turbollm launch openclaw` write a `turbollm` provider into that tool's own config file
+(preserving anything you already configured) and then launch it. If an existing config can't
+be parsed, TurboLLM refuses to overwrite it and prints how to add the provider by hand.
+`turbollm launch hermes` does the equivalent for [Hermes Agent](https://github.com/NousResearch/hermes-agent)
+via its own `hermes config set` command (its config is YAML, so we drive its CLI instead of
+hand-editing that file). The in-app **Developer** screen also shows copy-paste snippets for
+any OpenAI- or Anthropic-compatible tool (Open WebUI, Kilo Code, opencode, …).
 
 ---
 
@@ -397,6 +416,7 @@ turbollm --addr 0.0.0.0:6996    # bind all interfaces (LAN sharing)
 turbollm --stop                 # stop a running daemon (any terminal)
 turbollm launch claude          # start Claude Code (auto-loads a model if none is running)
 turbollm launch claude --model qwen3-8b   # load a specific model, then launch
+turbollm launch opencode        # wire opencode (or kilo / openclaw) to TurboLLM, then launch
 ```
 
 | Flag | Description |
@@ -410,7 +430,9 @@ turbollm launch claude --model qwen3-8b   # load a specific model, then launch
 
 `turbollm launch claude` also accepts `--model <key|name>` to load a specific model before
 launching; without it, an already-loaded model is used, or the last-used / first model is
-auto-loaded.
+auto-loaded. The same command works with `opencode`, `kilo`, and `openclaw` — each gets a
+`turbollm` provider merged into its own config file before it starts — and with `hermes`,
+which gets configured via its own `hermes config set` command instead.
 
 ---