Local AI Picks

Finde Open-Weight-KI-Modelle, die zu deiner Hardware passen, mit Gewichtsgrößen, Speicher und Quellen.

Gehosteter MCP-Server

npx add-mcp 'https://localaipicks.com/mcp'

Installiert in Claude Code, Codex, Cursor und mehr

Dokumentation

MCP server

AI assistants that support the Model Context Protocol can read the same model list you see on this site. It is free and needs no key or account.

Endpoint: https://localaipicks.com/mcp (Streamable HTTP).

Connect

Claude Code:

claude mcp add --transport http localaipicks https://localaipicks.com/mcp

Cursor, VS Code and other clients that take a JSON config:

{
  "mcpServers": {
    "localaipicks": { "url": "https://localaipicks.com/mcp" }
  }
}

In Claude or ChatGPT, add it as a custom connector with the same URL.

Tools

  • list_tasks: The tasks this site covers, with the number of current models and the page for each.
  • find_models: Current models for one task, in site order. Pass memory_gb (and device) to keep only models that fit.
  • get_model: One model: parameters, license, context length, weight options with file size and memory need, scores with source and date, test status.
  • compare_models: The same facts for 2 to 4 models side by side.

All tools are read-only. Answers carry the same numbers, sources and dates as the pages, plus a link to the page. There are no ratings or speed estimates. Each IP address can make 60 requests a minute.