Stoneforge

Stoneforge is an open-source, multi-agent orchestration tool designed to run teams of AI coding agents in parallel. It provides dependency-aware task dispatch, automated code merging using isolated git worktrees, and a real-time web dashboard. The platform is aimed at developers looking to manage and coordinate multiple AI coding workers simultaneously.

• Dependency-aware task dispatch that auto-assigns ready tasks to idle workers based on priorities and dependencies
• Automated code merging using isolated git worktrees to prevent merge conflicts when agents edit files
• Real-time web dashboard for monitoring live agent output, kanban task progress, merge status, and performance metrics
• Support for persistent long-lived workers handling multiple tasks as well as ephemeral short-lived workers spawned per task
• Merge stewards that check CI status, resolve conflicts, and handle integration
• Integration with external synchronization tools including Github, Linear, and Notion
• Orchestration support for AI coding tools like Claude Code, Codex, and OpenCode

Pros

• Prevents overlapping and duplicated work by automatically tracking task priorities and clearing dependencies before assigning jobs
• Eliminates common file conflict issues by running every worker in an isolated git worktree
• Provides deep visibility into parallel operations through a central web dashboard featuring live command center views and activity feeds

Cons

• Pricing is not published on the page and requires users to contact them for pricing
• Requires external AI coding agents like Claude Code, Codex, or OpenCode to function fully as an orchestration layer

Stoneforge fits developers and teams looking to run and coordinate multiple AI coding agents simultaneously without suffering from merge conflicts or visibility bottlenecks. It is poorly suited for solo developers who only use a single sequential prompt workflow or teams not utilizing AI coding agents like Claude Code, Codex, or OpenCode.

The product is open source, and the listing states you have to contact them for pricing.

Unlike traditional isolated terminal workflows where individual agent actions must be monitored separately, Stoneforge introduces a dedicated coordination daemon and merge stewards. It handles multi-agent scaling by combining a kanban command center with automated git worktree isolation.

Stoneforge is worth trying for developers and teams looking to scale up multi-agent coding workflows with automated merging and dashboard visibility. Because the platform is open source, users can inspect and deploy it, though pricing for commercial tiers or support must be requested directly. Teams relying heavily on parallel AI agents will find its dispatch and isolation features helpful, whereas single-agent users may find it overly complex.