Forge provides an open-source control plane designed to manage and automate AI agent operations using your local Claude Code instances. It acts as a central hub, allowing you to trigger AI agent work via webhooks from various sources like GitHub or Sentry, and then guides these tasks through a configurable multi-stage pipeline. The system ensures that all jobs, including stdout, stderr, tool calls, and diffs, are fully auditable and replayable, offering real-time streaming to a web dashboard. This self-hosted solution integrates with your own devices running Claude Code, ensuring that sensitive AI credentials remain local and secure.
Key Features
01Multi-project and multi-device management for AI agent operations
02Customizable multi-stage workflow pipeline for issues
031 GitHub stars
04Self-hosted AI agent orchestration control plane
05Real-time job monitoring with full audit trail and replayability
06Webhook-driven ingestion for issues from any source
