MCP server that sends one task to 3 models in parallel and returns their short answers for Claude to summarize. It supports two presets, `china` and `usa`, using different model sets for each.
From the registry: Three-model MCP server returning Grok, Gemini, and DeepSeek outputs for Claude.
$ curl -LsSf https://astral.sh/uv/install.sh | shPlease install the `model-council-mcp-server` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install uv** — Install the uv Python package manager/runner, which is used to launch the MCP server via uvx. Run: `curl -LsSf https://astral.sh/uv/install.sh | sh` (https://docs.astral.sh/uv/)
- **Install Python 3.11+** — Python 3.11 or newer is required to run the server via uvx.
- **Create a Polza AI API key** — Generate a POLZA_AI_API_KEY for authenticating requests to the Polza AI API. (https://polza.ai/dashboard/api-keys)
Canonical MCP server config (stdio transport):
- command: `uvx`
- args: ["model-council-mcp-server@latest"]
- required environment variables:
- `POLZA_AI_API_KEY`: API key for Polza AI. Required for the server to function. (example: `<your-api-key>`)
- optional environment variables:
- `POLZA_BASE_URL`: Base URL for the Polza AI API. Override only if you need a different API endpoint. (example: `<your-polza-base-url>`)
- `COUNCIL_CHINA_MOONSHOT_MODEL`: Override the Moonshot model ID used in the china preset. (example: `<your-moonshot-model-id>`)
- `COUNCIL_CHINA_QWEN_MODEL`: Override the Qwen model ID used in the china preset. (example: `<your-qwen-model-id>`)
- `COUNCIL_CHINA_DEEPSEEK_MODEL`: Override the DeepSeek model ID used in the china preset. (example: `<your-deepseek-model-id>`)
- `COUNCIL_USA_GEMINI_MODEL`: Override the Gemini model ID used in the usa preset. (example: `<your-gemini-model-id>`)
- `COUNCIL_USA_GROK_MODEL`: Override the Grok model ID used in the usa preset. (example: `<your-grok-model-id>`)
- `COUNCIL_USA_OPENAI_MODEL`: Override the OpenAI model ID used in the usa preset. (example: `<your-openai-model-id>`)
- `LOG_LEVEL`: Logging verbosity for the server. (example: `<your-log-level>`)
Note: Preferred install path is PyPI via uvx. The README also includes a local development path, but the portable config uses stdio with `uvx model-council-mcp-server@latest`.
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.POLZA_AI_API_KEYrequiredAPI key for Polza AI, used to call the models.AI orchestration with hive-mind swarms, neural networks, and 87 MCP tools for enterprise dev.
Persistent codebase knowledge graph. Survives session restarts and context compaction.
A powerful toolkit for coding, providing semantic retrieval and editing capabilities.