#1515th of 15 in MCP servers for developers

pal-mcp-server

MCP server that lets Claude Code, Codex CLI and Cursor call multiple LLMs

Stars
11.8k
License
custom license
Last commit
Dec 2025
Last release
Dec 2025
Language
Python
Stars, last 29 weeks+700
25 Mar 2026: 11,084 stars1 Apr 2026: 11,136 stars8 Apr 2026: 11,180 stars15 Apr 2026: 11,234 stars22 Apr 2026: 11,276 stars29 Apr 2026: 11,310 stars6 May 2026: 11,348 stars13 May 2026: 11,373 stars20 May 2026: 11,401 stars27 May 2026: 11,421 stars3 Jun 2026: 11,442 stars10 Jun 2026: 11,455 stars17 Jun 2026: 11,470 stars24 Jun 2026: 11,490 stars1 Jul 2026: 11,520 stars8 Jul 2026: 11,543 stars15 Jul 2026: 11,580 stars22 Jul 2026: 11,603 stars29 Jul 2026: 11,618 stars5 Aug 2026: 11,637 stars12 Aug 2026: 11,648 stars19 Aug 2026: 11,664 stars26 Aug 2026: 11,678 stars2 Sept 2026: 11,688 stars9 Sept 2026: 11,703 stars16 Sept 2026: 11,720 stars23 Sept 2026: 11,732 stars30 Sept 2026: 11,745 stars7 Oct 2026: 11,765 stars14 Oct 2026: 11,784 stars

11,084 on 25 Mar 202611,784 on 14 Oct 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

PAL is a Python MCP server that exposes tools such as chat, thinkdeep, planner, consensus, codereview, precommit and debug to a host CLI or IDE, and routes each call to a chosen model from Gemini, OpenAI, Azure, X.AI, OpenRouter, DIAL or Ollama. It keeps conversation threads across tools and models, and its clink tool launches other CLIs (Gemini CLI, Codex, Claude Code) as subagents with role prompts.

Who it is for: Developers using Claude Code, Codex or Cursor who want multi-model input

Strengths

  • Works with many hosts: Claude Code, Gemini CLI, Codex CLI, Qwen CLI, Cursor
  • Supports Gemini, OpenAI, Azure, X.AI, OpenRouter, DIAL and local Ollama models
  • Conversation threading carries context across models and after host context resets
  • clink spawns other CLIs as isolated subagents with roles like planner and codereviewer

Weaknesses

  • Needs at least one provider API key or a local Ollama setup
  • Each enabled tool adds to the context window; six tools are disabled by default
  • License is reported as NOASSERTION, so terms need manual checking
  • Last commit was 2025-12-15, so recent model releases may lack support

What it needs

  • GPU optional
  • Docker + Compose
  • Needs Python 3.10+, Git, uv, Provider API key (Gemini, OpenAI, OpenRouter, Azure, X.AI, DIAL) or Ollama
  • Models: Gemini, OpenAI, Anthropic, Grok, Azure OpenAI

Also in MCP servers for developers

See all 15
Also in MCP servers for developers
RankProjectScore
1gitnexusCode knowledge graph exposed to coding agents through MCP tools Live demo ↗ (opens in a new tab)47.9k starscustom license89 out of 100
2codegraphLocal code knowledge graph served to coding agents over MCP74.1k starsMIT81 out of 100
3chrome-devtools-mcpMCP server that lets coding agents drive and debug a live Chrome53.6k starsApache-2.070 out of 100
#4codebase-memory-mcpLocal MCP server that indexes code into a queryable knowledge graph46.6k starsMIT69 out of 100
#5github-mcp-serverMCP server exposing GitHub repos, issues, PRs and Actions to AI tools33.6k starsMIT68 out of 100