PAL MCP Server
BeehiveInnovations/pal-mcp-serverReviewed by AI
Checked by RepoPicks Radar on
What is PAL MCP Server?
PAL MCP (formerly Zen MCP) is an MCP server you plug into a coding assistant such as Claude Code, Codex CLI, Gemini CLI or Cursor. From inside your usual chat you can send a task to other models — Gemini, OpenAI, Grok, Azure, OpenRouter, DIAL or a local Ollama model — and it keeps the conversation going between them: code review by two models, a planner, a debug walkthrough, a pre-commit check, a consensus debate. You install it with git and uv, put at least one model key in a .env file and point your assistant's config at it.
How easy is it to install PAL MCP Server?
C · install 1 or 2 tools · macOS
- Windows D
- macOS C
- Linux C
On Mac and Linux it is a C: the prerequisites are Python 3.10+, Git and uv, then you clone the repository and run ./run-server.sh, which writes the settings for Claude Code, Gemini CLI, Codex CLI and Qwen CLI by itself. Windows is a D because the README's commands are bash and the project keeps a separate "WSL Setup - Windows users" guide, so you first set up the Linux subsystem.
PAL MCP Server on GitHub
- Stars
- 11,838on Nov 10, 2026
- Downloads
- 1,77365 this week
- Growth
- +14stars in 6 days
Stars over time
Facts
- Runs on
- Windows, macOS, Linux
- You need to install
- Python, Other tools
- To build from source
- Python, Docker
- Disk space
- 100 MB–300 MB · estimate
- Accounts or keys
- At least one model provider: Google Gemini, OpenAI, Azure OpenAI, X.AI Grok, DIAL or OpenRouter, A local Ollama or LM Studio model instead of a paid key
- Cost
- Free, API costs extra · The server is free (Apache 2.0), but it only does something with a model key you pay for; the README also lists Ollama local models as a free option.
- License
- Apache-2.0 (Permissive)
- Can replace
- CodeRabbit
- Last change on GitHub
- Dec 15, 2025
PAL MCP Server risk signals (5)
- Attentioninstall option runs the latest code from git
Install "Option B" puts this in your assistant's settings: `uvx --from git+https://github.com/BeehiveInnovations/pal-mcp-server.git pal-mcp-server` — no version pinned, so every start downloads and runs whatever is on the repository's main branch at that moment, with your API keys in the same config block.
AI review
- Infoyour files go to several ai companies
README: "Smart file handling - Auto-expand directories" and, in the code review example, step 6 "Shares the relevant files, findings, etc with Gemini Pro" — the code you are working on is sent to whichever providers you enable.
AI review
- Infono updates for eleven months
Last commit and last release are both 2025-12-15, 325 days before the check, with 153 open issues, while the model lists in conf/ and .env.example name versions that keep changing.
AI review
- Inforenamed project
The README says "Formerly known as Zen MCP" and links docs/name-change.md, so guides and configs found elsewhere may still use the old name.
AI review
- Infohalf the tools are off by default
README Tool Configuration: "DISABLED_TOOLS=analyze,refactor,testgen,secaudit,docgen,tracer" — those six only work after you edit the list, because "Each tool adds to context window usage".
AI review
Evidence is shown in its original language.
What Radar could not check about PAL MCP Server
- How much the calls to Gemini, OpenAI or Grok cost in a normal day of work
- Whether run-server.ps1 really sets everything up on Windows without the Linux subsystem
- Whether the model names in conf/ still exist at the providers, after almost a year without updates
- How well the local Ollama option works compared with the paid models
- What exactly each tool sends to each provider on a real project