Maige
Open-source AI infrastructure for running natural language workflows on your GitHub codebase, automating issue labeling, assignment, code review, and more.
At a Glance
About Maige
Maige is an open-source, AI-powered codebase copilot built by Rubric Labs that lets teams automate GitHub repository maintenance using plain English instructions. It connects to your repository via a GitHub App, creates codebase embeddings, and runs a webhook-driven workflow engine that responds to issues and pull requests in real time. The project is currently in alpha and is licensed under AGPL-3.0.
What It Is
Maige sits between your GitHub repository and an AI layer, translating natural language rules into automated GitHub actions. Instead of writing custom bots or GitHub Actions YAML, maintainers write plain-text instructions like "always assign UI-related issues to @username" or "label .env PRs as 'needs-approval' unless opened by @maintainer." Maige interprets these rules and executes them automatically whenever issues or PRs are opened.
How the Workflow Works
Getting started involves three steps:
- Connect your repo: Maige installs a webhook, generates embeddings of your entire codebase, and provisions a sandbox environment.
- Write your rules: Describe what should happen when issues and PRs are opened. Rules can trigger labeling, assignment, commenting, code review, or even simple code execution.
- Monitor runs: A dashboard lets maintainers watch workflow executions and provide feedback.
The system works flexibly with the GitHub API, meaning it can perform any action a human could take through the GitHub UI.
Capabilities
Maige supports a range of automated actions out of the box:
- Auto-labeling: Classify and tag issues and PRs based on content
- Auto-assignment: Route issues to the right contributors
- Auto-comments: Post contextual responses to new issues or PRs
- Code review: Analyze incoming PRs against project guidelines (e.g., CONTRIBUTING.md)
- Code generation: Run simple code snippets in a sandboxed environment
- Custom instructions: Any freeform natural language rule that maps to a GitHub API action
Open-Source Deployment Model
The core platform is published on GitHub under the AGPL-3.0 license, making the source code freely available to inspect, modify, and self-host. The README notes that self-hosting is not yet fully supported and may involve rough edges, but provides setup instructions using Bun, Vercel, and ngrok for local development. The GitHub App integration requires configuring webhooks, permissions for Issues and Pull Requests, and a private key. Rubric Labs also operates a hosted version at maige.app for teams that prefer a managed deployment.
Current Status
The project is labeled alpha on the homepage and was created in March 2023, with the repository last updated in mid-2026. The homepage notes the hosted service is used by over 4,300 repositories, citing projects such as Documenso, Nuxt, Highlight.io, Cal.com, and Trigger.dev as examples (per vendor-published homepage claims). An Enterprise plan is listed as "coming soon." The GitHub repository has 102 stars and 11 forks as of the last recorded update.
Community Discussions
Be the first to start a conversation about Maige
Share your experience with Maige, ask questions, or help others learn from your insights.
Pricing
Open Source
Self-host the AGPL-3.0 licensed codebase for free.
- Auto-labeling
- Auto-assignment
- Auto-comments
- Custom instructions
- Code review
Standard Plan
Hosted plan with first 30 issues free.
- Auto-labeling
- Auto-assignment
- Auto-comments
- Custom instructions
- Code review
- Code generation
- First 30 issues free
Enterprise
Best for large teams. Coming soon.
- Large team support
Capabilities
Key Features
- Natural language workflow rules
- Auto-labeling of issues and PRs
- Auto-assignment of contributors
- Auto-comments on issues and PRs
- AI-powered code review
- Code generation in sandbox
- Codebase embeddings
- GitHub webhook integration
- Custom instructions via plain text
- Dashboard for monitoring runs
