Bernstein
Audit-grade multi-agent orchestration for CLI coding agents (Claude Code, Codex, Gemini CLI, +40 more). HMAC-chained audit log, signed agent cards, per-artefact lineage, air-gap deploy. The orchestrator your compliance team will sign off on. https://
What is Bernstein?
Bernstein is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to audit-grade multi-agent orchestration for cli coding agents (claude code, codex, gemini cli, +40 more). hmac-chained audit log, signed agent cards, per-artefact lineage, air-gap deploy. the orchestrat...
Audit-grade multi-agent orchestration for CLI coding agents (Claude Code, Codex, Gemini CLI, +40 more). HMAC-chained audit log, signed agent cards, per-artefact lineage, air-gap deploy. The orchestrator your compliance team will sign off on. https://
This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Audit-grade multi-agent orchestration for CLI coding agents
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx bernsteinConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Bernstein
Bernstein is an audit-grade multi-agent orchestration platform for CLI coding agents such as Claude Code, Codex, Gemini CLI, and 40+ others. It runs agents in parallel inside isolated git worktrees with a deterministic Python scheduler, captures a tamper-evident HMAC-SHA256 audit chain per RFC 2104, signs agent cards with Ed25519, and records per-artefact lineage including producer, prompt SHA, model, and cost. Compliance and security teams use it when they need a full chain of custody for AI-generated code, while engineering teams use it to fan out the same task across multiple agents and accept the best result.
Prerequisites
- Python 3.12+ installed
- Git installed (required for worktree isolation)
- At least one CLI coding agent installed (e.g., Claude Code, Codex CLI, Gemini CLI)
- An API key for the LLM provider used by your chosen agent (e.g., ANTHROPIC_API_KEY for Claude Code)
- An MCP-compatible client or terminal to issue bernstein commands
Install Bernstein
Install via pip, pipx, or your preferred method. The pipx method is recommended for isolated CLI tool installation.
pip install bernstein
# or
pipx install bernstein
# or
uv tool install bernstein
# or (macOS)
brew tap chernistry/tap && brew install bernsteinInitialise a project
Run 'bernstein init' inside your git repository to create a bernstein.yaml configuration file where you specify which agents to use, sandbox backends, and the internal LLM for orchestration routing.
cd /path/to/your/repo
bernstein initSet required environment variables
Export API keys for your LLM provider(s). For Claude Code as the coding agent, set ANTHROPIC_API_KEY. For E2B sandboxing, set E2B_API_KEY.
export ANTHROPIC_API_KEY=sk-ant-...
# Optional extras:
export E2B_API_KEY=e2b_...Run a task
Use 'bernstein run' with a goal flag to dispatch the task to your configured agents in parallel. Bernstein will coordinate execution, run tests, and produce an audit log.
bernstein run -g "Add rate limiting to the /api/users endpoint"Review the audit log and artefacts
After the run, inspect the HMAC-chained audit log and per-artefact lineage records. Bernstein records the producer agent, model, prompt SHA, and cost for every generated file.
bernstein audit log
bernstein audit show --artefact path/to/file.pyCreate a pull request automatically
Bernstein can open a GitHub pull request from the best-scoring agent run output.
bernstein prBernstein Examples
Client configuration
MCP server config for Claude Desktop to expose Bernstein's orchestration tools. The npx wrapper is also available.
{
"mcpServers": {
"bernstein": {
"command": "npx",
"args": ["bernstein-orchestrator"]
}
}
}Prompts to try
CLI commands and workflow invocations representative of typical Bernstein usage.
- "bernstein run -g 'Fix the failing test in tests/test_auth.py'"
- "bernstein workflow run idea-to-pr -g 'Add JWT authentication'"
- "bernstein from-ticket https://github.com/org/repo/issues/42"
- "bernstein autofix # monitor and repair failed PRs"
- "bernstein chat serve --platform=slack # drive runs from Slack"Troubleshooting Bernstein
bernstein init fails with 'not a git repository'
Bernstein requires a git repository for worktree isolation. Run 'git init' in the project directory first, or run bernstein from within an existing git repo.
Agent runs start but produce no output
Confirm the CLI coding agent is installed and on your PATH (e.g., 'claude --version' or 'codex --version'). Also verify the relevant API key environment variable is exported and valid.
Audit log verification fails
The HMAC chain will break if log files are modified manually. Do not edit files under the .bernstein/ audit directory. If corruption is suspected, run 'bernstein audit verify' to identify the first tampered entry.
Frequently Asked Questions about Bernstein
What is Bernstein?
Bernstein is a Model Context Protocol (MCP) server that audit-grade multi-agent orchestration for cli coding agents (claude code, codex, gemini cli, +40 more). hmac-chained audit log, signed agent cards, per-artefact lineage, air-gap deploy. the orchestrator your compliance team will sign off on. https:// It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Bernstein?
Follow the installation instructions on the Bernstein GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Bernstein?
Bernstein works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Bernstein free to use?
Yes, Bernstein is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.
Bernstein Alternatives — Similar Coding Agents Servers
Looking for alternatives to Bernstein? Here are other popular coding agents servers you can use with Claude, Cursor, and VS Code.
Dify
★ 142.2kProduction-ready platform for agentic workflow development.
Ruflo
★ 54.0k🌊 The leading agent orchestration platform for Claude. Deploy intelligent multi-agent swarms, coordinate autonomous workflows, and build conversational AI systems. Features enterprise-grade architecture, self-learning swarm intelligence, RAG integrat
Goose
★ 45.7kan open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
Antigravity Awesome Skills
★ 38.3kInstallable GitHub library of 1,400+ agentic skills for Claude Code, Cursor, Codex CLI, Gemini CLI, Antigravity, and more. Includes installer CLI, bundles, workflows, and official/community skill collections.
AgentScope
★ 25.5kBuild and run agents you can see, understand and trust.
Serena
★ 24.5kA coding agent toolkit that provides IDE-like semantic code retrieval and editing tools, enabling LLMs to efficiently navigate and modify codebases using symbol-level operations instead of basic file reading and string replacements.
Browse More Coding Agents MCP Servers
Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Bernstein in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Bernstein?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.