Skip to content
MCP server

Bernstein

Open-source scheduler that runs coding agents in parallel under policy-as-code rules

Runs locallyStatus and history
Pricing
Open source
Platforms
CLI, Linux, macOS
Runs
Locally
Version
3.24.1
Listing updated
3 days ago

About Bernstein

Bernstein is an open-source MCP server for developer tools. It runs locally on your machine and is started by your AI client.

Bernstein sits above your command-line coding agents rather than being one itself. A deterministic Python scheduler, with no model in the coordination loop, hands tasks to agents in parallel, isolates each in its own git worktree, checks lint, types and tests per diff, and merges only what passes. Because the plan is deterministic, it replays to the same task graph.

You write the policy: who may do what, what needs approval, what must be recorded. Bernstein enforces it and produces a signed run receipt that can be verified offline. The site says it works around Claude Code, Codex, Gemini CLI and 40 or more other adapters, and runs on a laptop, on-prem or air-gapped.

The README is blunt about maturity: it is beta, solo-maintained, and version numbers count releases rather than stability, so pinning a version is advised. It installs from PyPI or as an OCI image.

Install

Run in a terminal

claude mcp add bernstein -- uvx bernstein

Add to claude_desktop_config.json

{
  "mcpServers": {
    "bernstein": {
      "command": "uvx",
      "args": [
        "bernstein"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "bernstein": {
      "command": "uvx",
      "args": [
        "bernstein"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "bernstein": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "bernstein"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "bernstein": {
      "command": "uvx",
      "args": [
        "bernstein"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "bernstein": {
      "command": "uvx",
      "args": [
        "bernstein"
      ]
    }
  }
}

What it does

Deterministic scheduling
No model in the coordination loop, so a plan replays to the same graph.
Worktree isolation
Each agent works in its own git worktree.
Per-diff verification
Runs lint, type and test checks before anything merges.
Policy as code
Approvals and records are set by policy you write.
Signed run receipts
Receipts verify offline.
Air-gap friendly
Runs on a laptop, on-prem or without internet access.

Good for

  • Run several coding agents on one repository at once
  • Require approvals before agents change sensitive paths
  • Keep an auditable record of what each agent did
  • Operate agent workflows in an air-gapped environment

Pricing

Open source

Open source; sponsorship welcome

Prices change. Check bernstein.run before you buy.

Bernstein FAQ

Is it production-ready?

The README calls it beta and solo-maintained, and advises pinning the version.

How is it installed?

From PyPI as bernstein, or as an OCI image on ghcr.io.

Which agents does it work with?

The site names Claude Code, Codex and Gemini CLI plus 40 or more adapters.

Does an LLM decide the task order?

No. The scheduler is deterministic with no model in the coordination loop.

Is Bernstein free?

Yes. Bernstein is open source, so you can use and self-host it for free.

How do I add Bernstein to Claude Code?

Run this command in your terminal: claude mcp add bernstein -- uvx bernstein

What are the best alternatives to Bernstein?

The closest alternatives to Bernstein are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

Alternatives to Bernstein

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  4. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  5. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  6. Ouroboros

    MCP server

    Spec-first layer that pins requirements and verifies results of AI coding agent work

    Runs locally
    • Open source
    • 6.4K GitHub stars
    • Developer tools