Consult other LLMs as tools โ ask them, compare their answers, synthesize a conclusion.
Connect Runs on your machine
claude mcp add model-council -e COUNCIL_MODELS=<COUNCIL_MODELS> -e COUNCIL_CONFIG=<COUNCIL_CONFIG> -- uvx model-council-mcp==0.14.0
{
"mcpServers": {
"model-council": {
"command": "uvx",
"args": [
"model-council-mcp==0.14.0"
],
"env": {
"COUNCIL_MODELS": "<COUNCIL_MODELS>",
"COUNCIL_CONFIG": "<COUNCIL_CONFIG>"
}
}
}
}
{
"mcpServers": {
"model-council": {
"command": "uvx",
"args": [
"model-council-mcp==0.14.0"
],
"env": {
"COUNCIL_MODELS": "<COUNCIL_MODELS>",
"COUNCIL_CONFIG": "<COUNCIL_CONFIG>"
}
}
}
}
{
"servers": {
"model-council": {
"type": "stdio",
"command": "uvx",
"args": [
"model-council-mcp==0.14.0"
],
"env": {
"COUNCIL_MODELS": "<COUNCIL_MODELS>",
"COUNCIL_CONFIG": "<COUNCIL_CONFIG>"
}
}
}
}
Replace the values in angle brackets with your own. Keys stay on your machine; we never see them.
Tools
This server runs locally, so its tools are listed by your MCP client once it starts.
Environment variables
| Name | Description |
|---|---|
COUNCIL_MODELSOptional | Comma-separated short ids for the models to seat, e.g. 'gpt5,glm'. Each id then takes its own <ID>_BASE_URL, <ID>_API_KEY, <ID>_MODEL and optional <ID>_FORMAT (openai or anthropic). Omit to use a config file instead. |
COUNCIL_CONFIGOptional | Path to a JSON config file, for rosters large enough that several models share one endpoint. Defaults to ~/.config/model-council/config.json when that exists. |
Details
| Package | [email protected] (PYPI, stdio) |
| First published | 24 Aug 2026 |
| Last updated | 24 Aug 2026 |
| Registry name | io.github.Totti0135/model-council |
More ai and media servers See all
Listing data from the official MCP registry. Status and tools from NestDaddy's daily check. Browse all MCP servers.
