Get This Tool
Gorchestra
Pricing
- Model
- Free
Summary
Running three AI coding agents in parallel means three disconnected terminals, no shared context, and a session you cannot replay after the first crash — gorchestra exists to hold all of that together.
Gorchestra is a locally-run Go server that wraps multiple Codex and Claude coding sessions into a single interface, streaming every event in real time and persisting the full history to SQLite. You can inspect live file changes and git state mid-run, queue follow-up messages without interrupting the agent, and return to the exact same session state after a disconnect or restart. The tool is MIT-licensed, self-hosted, and carries no hosted backend. The ceiling appears when your workflow demands coordinating agents across machines or teams — everything here is designed around a single developer's local environment, not a shared multi-user setup.
Bottom line: Pick gorchestra when you are one developer running parallel Claude or Codex sessions locally and need durable replay; look elsewhere the moment your workflow requires shared access across a team or remote execution.
Community Performance Report Card
No community ratings yet. Be the first to rate this tool!
Community Benchmarks Community
Sign in to submit a benchmarkNo community benchmarks yet. Be the first to share a real-world data point.
Pros
Sign in to edit- Full session history persisted to a local SQLite file, so a disconnect or restart does not erase the record of what the agent did — you replay from the stored event log rather than starting over.
- Real-time file and git state inspection during agent runs, which means you catch a runaway edit or a bad branch decision while the session is still live instead of discovering it in a post-mortem.
- Desktop and mobile web UI over a local server, so you can monitor or queue messages from a phone without keeping a terminal window open on the same machine.
- MIT license with no hosted component or paid tier, which means there is no vendor dependency, no data leaving your machine, and no pricing gate blocking any feature.
- Parallel session management across Codex and Claude agents from a single interface, so switching between tasks does not mean context-switching between unrelated terminal windows.
Cons
Sign in to edit- The architecture is explicitly single-machine and single-user: there is no shared session model, no role-based access, and no way for two developers to observe or hand off the same running session. A team that needs that coordination hits a wall at the first handoff and moves to a hosted platform with multi-user primitives.
- No API surface is exposed, so gorchestra cannot be triggered or queried by an external system — CI pipelines, Slack bots, or custom dashboards cannot integrate with it. Teams building any kind of automated trigger around their agent runs have to instrument the agents themselves rather than the orchestration layer.
- Support for agents is scoped to Codex and Claude per the repository description. Teams standardized on a different coding agent or building a custom tool-use loop find no documented extension point and will need a different runner.
Community Reviews
Sign in to write a reviewNo reviews yet. Be the first to share your experience.
About
- Platforms
- Desktop, Mobile, Local server
- API Available
- No
- Self-Hosted
- Yes
- Last Updated
- 2026-06-22T08:15:53.281Z
Best For
Who it's for
- Developers coordinating multiple AI coding agents locally
- Teams needing durable, inspectable agent execution history
- Users wanting a unified desktop or mobile interface for agent workflows
What it does well
- Running parallel AI coding sessions across agents
- Inspecting live file changes and git state during agent runs
- Replaying or resuming sessions after disconnects or restarts
Integrations
Discussion Community
Sign in to commentNo discussion yet. Sign in to start the conversation.
Compare Gorchestra
Spotted incorrect or missing data? Join our community of contributors.
Sign Up to ContributeCommunity Notes & Tips Community
Sign in to contributeBe the first to contribute. General notes, observations, gotchas, and tips from people who use this tool day-to-day.
Frequently Asked Questions
- Is Gorchestra free?
- Yes — Gorchestra is fully free to use. There is no paid tier.
- Is Gorchestra open source?
- Yes. Gorchestra is open source.
- Can I self-host Gorchestra?
- Yes. Gorchestra supports self-hosting on your own infrastructure.
- What platforms does Gorchestra support?
- Gorchestra is available on: Desktop, Mobile, Local server.
Hours Saved & ROI Stories Community
Sign in to contributeBe the first to contribute. Concrete time/cost savings, with context. e.g. "Cut my code review backlog from 4h to 45m per week."
Curated lists that include this category
Gorchestra is a Go-based server with a local web UI that acts as a control room for AI coding agents. The core workflow is straightforward: you start sessions against Codex or Claude, gorchestra streams every tool call and file event to a live feed, and SQLite records the full history so nothing is lost across refreshes, reconnects, or restarts. You can queue follow-up messages, upload screenshots for context, and switch into planning mode without tearing down the session.
The standout capability is durable execution history. Most terminal-based agent runners treat a disconnect as a hard stop. Gorchestra stores the entire event sequence locally, so you can replay a session, inspect what the agent touched in the file tree, and check git state at any point in the run — a materially different debugging surface compared to reading a scrollback buffer after the fact.
The tool fits a solo developer who runs multiple long-horizon coding tasks in parallel and wants a desktop or mobile interface that survives network interruptions. The hard limit is its single-machine, single-user architecture. There is no multi-user access model, no remote agent execution, and no API surface for integrating gorchestra into a CI pipeline or shared team tooling. Teams that need those capabilities will reach for a hosted orchestration layer and will not find a migration path here.
Installation is handled via Homebrew packaging included in the repository. The server is entirely local — the vendor describes no telemetry, no cloud dependency, and no paid tier. The SQLite persistence layer means the event store is a plain file you can inspect, back up, or delete without any vendor involvement.