Install Sezzlee Local LLM in ECA
What you need first
- ECA, installed.
- Node.js, which
gives the
npxcommand the server starts with. - Your own value for
YOUR_SEZZLEE_LLM_MODEL.
Steps
- Open
~/.codex/config.toml, or create it. The Codex CLI, its IDE extension and the ChatGPT desktop app share this file. - Add this table at the end of the file.
{
"mcpServers": {
"llm": {
"command": "npx",
"args": ["-y", "@sezzlee/[email protected]"],
"env": { "SEZZLEE_LLM_MODEL": "YOUR_SEZZLEE_LLM_MODEL" }
}
}
} - Replace
YOUR_SEZZLEE_LLM_MODELwith your own value. - Save the file and restart ECA.
Check it worked
Ask ECA: List the tools of llm.
If it does not start
- spawn npx ENOENT
- Node.js is not on the PATH that ECA sees. Install it, or write the
full path of
npxin the config. - The server starts, but its tools fail
- A
YOUR_…placeholder is still in the config, or its value is wrong. - Nothing changes after a restart
- The file is not valid JSON any more. Look for a missing comma where the block was added.