Vestige

Please install the `vestige` MCP server into my current AI client (that's you).

Required prerequisites (do these first if not already done):
- **Install Vestige globally via npm** — Install the Vestige MCP server package so the `vestige-mcp` command is available on your PATH. Run: `npm install -g vestige-mcp-server@latest`

Optional prerequisites:
- Intel Mac only: install ONNX Runtime — On macOS Intel, Vestige requires a Homebrew-installed ONNX Runtime and the ORT_DYLIB_PATH environment variable pointing to the dynamic library. Run: `brew install onnxruntime` (docs/INSTALL-INTEL-MAC.md)

Canonical MCP server config (stdio transport):
- command: `vestige-mcp`
- args: []
- optional environment variables:
  - `ORT_DYLIB_PATH`: macOS Intel only. Path to the ONNX Runtime dynamic library required by Vestige's Intel Mac build. (example: `<path-to-libonnxruntime.dylib>`)
  - `VESTIGE_DASHBOARD_ENABLED`: Set to enable the local dashboard alongside the MCP server. (example: `true`)
  - `FASTEMBED_CACHE_PATH`: Override the cache directory used for downloaded embedding models. (example: `<your-fastembed-cache-path>`)
  - `HTTPS_PROXY`: Proxy to use when downloading the embedding model on first run, if needed. (example: `<your-proxy-host:port>`)

Note: Vestige is an MCP stdio server launched with `vestige-mcp`. First run downloads an embedding model (~130MB) and then works offline. The README also mentions a local dashboard available separately via `vestige dashboard` or by setting `VESTIGE_DASHBOARD_ENABLED=true` when running the MCP server.

Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.