Devswarm
High-performance MCP server, code graph engine & evolutionary algorithm platform in Zig. 33 tools: GitHub project management, agent swarm orchestration, iterative review-fix loops, blast radius analysis, and code navigation via Model Context Protocol
What is Devswarm?
Devswarm is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to high-performance mcp server, code graph engine & evolutionary algorithm platform in zig. 33 tools: github project management, agent swarm orchestration, iterative review-fix loops, blast radius analys...
High-performance MCP server, code graph engine & evolutionary algorithm platform in Zig. 33 tools: GitHub project management, agent swarm orchestration, iterative review-fix loops, blast radius analysis, and code navigation via Model Context Protocol
This server falls under the Coding Agents and Version Control categories on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- High-performance MCP server, code graph engine & evolutionar
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx devswarmConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Devswarm
Devswarm is a high-performance MCP server written in Zig that provides 37 tools for orchestrating AI agent swarms, managing GitHub projects, analyzing code structure, and running iterative review-fix loops against a local repository. It exposes tools for decomposing features into tasks, running parallel agent swarms (via claude or codex CLIs), performing blast radius analysis to understand the impact of changes, navigating code with caller/dependent graphs, and managing the full GitHub workflow (issues, branches, PRs, merges) from within an AI conversation. Engineering teams use it to automate multi-step development workflows, accelerate code review cycles, and coordinate multiple AI agents working on the same codebase.
Prerequisites
- Zig 0.15.x (if building from source), or a pre-built devswarm binary downloaded from GitHub Releases
- Git installed and the target repository accessible on disk
- The claude CLI (from @anthropic-ai/claude-code) and/or the codex CLI on your PATH for agent swarm execution
- GitHub CLI (gh) configured if using GitHub project management tools
- An MCP-compatible client such as Claude Code, Cursor IDE, or Amp
Download the devswarm binary
Download the latest pre-built binary for your platform from the GitHub Releases page. Make it executable and move it to a directory on your PATH.
# Example for macOS (check releases page for your platform binary)
chmod +x ./devswarm
sudo mv ./devswarm /usr/local/bin/devswarm
# Or build from source with Zig 0.15.x
git clone https://github.com/justrach/devswarm.git
cd devswarm && zig buildConfigure the server in Claude Code
Add devswarm to your ~/.claude.json MCP servers configuration, pointing REPO_PATH to the repository you want to work with.
{
"mcpServers": {
"devswarm": {
"command": "/usr/local/bin/devswarm",
"args": ["--mcp"],
"env": {
"REPO_PATH": "/absolute/path/to/your/repo"
}
}
}
}Verify tool availability
In your MCP client, run the /mcp command or check the tools list to confirm all 37 devswarm tools are registered and available.
Disable telemetry (optional)
If you prefer not to send anonymous usage data, set the DEVSWARM_TELEMETRY environment variable to false in your config.
{
"mcpServers": {
"devswarm": {
"command": "/usr/local/bin/devswarm",
"args": ["--mcp"],
"env": {
"REPO_PATH": "/absolute/path/to/your/repo",
"DEVSWARM_TELEMETRY": "false"
}
}
}
}Run your first agent task
Ask your AI assistant to use devswarm to decompose a feature or run a review-fix loop on a specific file or function.
Devswarm Examples
Client configuration
Example ~/.claude.json MCP configuration for devswarm with REPO_PATH set.
{
"mcpServers": {
"devswarm": {
"command": "/usr/local/bin/devswarm",
"args": ["--mcp"],
"env": {
"REPO_PATH": "/absolute/path/to/your/repo",
"DEVSWARM_TELEMETRY": "false"
}
}
}
}Prompts to try
After connecting, use these prompts to orchestrate agent workflows with devswarm.
- "Decompose the feature 'add OAuth2 login' into GitHub issues and prioritize them"
- "Run a swarm of 5 agents to audit the entire auth module for security issues"
- "Run a review-fix loop on src/queue.ts for memory leaks with up to 3 iterations"
- "Analyze the blast radius of changing the database connection pool settings"
- "Find all callers of the `processPayment` function and list files that would be affected by changes to its signature"
- "Create a branch, commit the current staged changes with context, and open a pull request against main"Troubleshooting Devswarm
Only some tools appear (fewer than 37) after connecting
Ensure REPO_PATH points to a valid Git repository directory. Some tools (git, PR, and issue tools) require the repository to have a remote configured. Run `git remote -v` in REPO_PATH to verify.
run_swarm or run_agent fails with 'claude not found'
Install the Claude CLI globally: `npm install -g @anthropic-ai/claude-code`. Verify it is on your PATH by running `claude --version` in a terminal. If using codex agents, install the codex CLI similarly.
devswarm binary crashes or outputs nothing on startup
Confirm the binary matches your OS and architecture (check the GitHub Releases page for the correct asset). If building from source, verify you are using Zig 0.15.x (`zig version`) as other versions may be incompatible.
Frequently Asked Questions about Devswarm
What is Devswarm?
Devswarm is a Model Context Protocol (MCP) server that high-performance mcp server, code graph engine & evolutionary algorithm platform in zig. 33 tools: github project management, agent swarm orchestration, iterative review-fix loops, blast radius analysis, and code navigation via model context protocol It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Devswarm?
Follow the installation instructions on the Devswarm GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Devswarm?
Devswarm works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Devswarm free to use?
Yes, Devswarm is open source and available under the AGPL-3.0 license. You can use it freely in both personal and commercial projects.
Devswarm Alternatives — Similar Coding Agents Servers
Looking for alternatives to Devswarm? 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 Devswarm 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 Devswarm?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.