You run models in LM Studio. bionic-mcp turns them into tools Claude Code (or any MCP client) can call, so you can hand the cheap and bulky work to your own machine. Free, local, private, nothing leaves your Mac.
You, in Claude Code: "classify these 50 log lines by severity"
↓
runs on your hardware, 0 API tokens
git clone https://github.com/turner-moore/bionic-mcp.git && cd bionic-mcp && npm install && \
claude mcp add -s user bionic -- node "$(pwd)/index.js"Then start LM Studio's local server (that's where the models are):
~/.lmstudio/bin/lms server startRegistering with a different MCP client
{
"mcpServers": {
"bionic": {
"command": "node",
"args": ["/absolute/path/to/bionic-mcp/index.js"],
"env": { "BIONIC_BASE_URL": "http://127.0.0.1:1234" }
}
}
}Two things would eventually bite you. This server fixes both:
Important
No silent model swaps. Ask LM Studio for a model that isn't loaded and its API quietly answers with whatever is loaded, so you think you're on a 14B and you're on a 3B. bionic-mcp refuses an unknown model id up front and checks response.model on every reply.
No surprise cold loads. On 16GB only one model stays resident, and switching costs 8-30s. model:"auto" uses whatever's already loaded, so a quick call stays quick.
| Tool | Parameters | Does |
|---|---|---|
respond |
prompt, model (default "auto"), system, temperature, json_schema |
Run a prompt on a local model. json_schema forces structured JSON out. |
models |
none | List installed models with live load state and a capability card (quality, speed, context). |
Note
Each prompt stands on its own. The local model can't see your Claude session, files, or memory, so inline every input and instruction it needs.
Bonus: a tiered on-device + LM Studio router (local.js)
local.js is a second, optional server. It tries a faster on-device backend first (caix, Apple Core AI on the Neural Engine) and falls back to LM Studio when that's down.
claude mcp add -s user local -- node "$(pwd)/local.js"- Node 18+ (for built-in
fetch) - LM Studio with its local server running
- The capability card lives in
models.json, edit it to match your own installed models.