Repolode
Submit
openai

openai/symphony

Turns project work into isolated, autonomous coding-agent runs so teams manage tasks instead of supervising agents directly.

View on GitHubWebsite
Stars
27,956
Forks
2,890
Open issues
0
Last commit
2 months ago

What is it?

Symphony is an orchestration layer for coding agents. It watches a work tracker such as a Linear board, picks up tasks, and starts an isolated agent run for each one. The agents carry the task to completion and produce proof of work, including CI status, pull request review feedback, complexity analysis, and walkthrough videos. Once a change is accepted, the agent lands the pull request.

The main idea is a shift in what engineers supervise. Instead of watching individual coding sessions, a team manages the queue of work and reviews the evidence agents produce. The project describes itself as best suited to codebases already set up for agent-driven development.

The project offers a written specification that can be used to build an implementation in any language, plus an experimental Elixir reference implementation. It is released as an engineering preview intended for testing in trusted environments.

Key features

  • 01Monitors a work tracker such as Linear and picks up tasks automatically.
  • 02Runs each task in an isolated agent session.
  • 03Produces proof of work such as CI status, review feedback and walkthrough videos.
  • 04Lands pull requests once changes are accepted.
  • 05Provides a language-neutral spec plus an Elixir reference implementation.

Use cases

  • A team wants engineers to manage a backlog of implementation tasks rather than babysit coding agents.
  • Experimenting with autonomous agent workflows in a trusted, non-production repository.
  • Building an agent orchestrator in a different language from the provided spec.

Quick start

  1. 1

    Read the Elixir reference implementation setup guide in elixir/README.md

  2. 2

    Or ask your coding agent to implement Symphony from the SPEC.md specification

Things to know

  • Symphony is described as a low-key engineering preview meant for testing in trusted environments.
  • The Elixir reference implementation is experimental.
  • It works best in codebases that have adopted harness engineering.

In collections

Open-source AI agent frameworks & appsAI coding assistants & agents

Topics

#ai-agents#coding-agent#orchestration#workflow-automation#linear#codex#elixir

FAQ

What is symphony?
Turns project work into isolated, autonomous coding-agent runs so teams manage tasks instead of supervising agents directly. Symphony is an orchestration layer for coding agents. It watches a work tracker such as a Linear board, picks up tasks, and starts an isolated agent run for each one. The agents carry the task to completion and produce proof of work, including CI status, pull request review feedback, complexity analysis, and walkthrough videos. Once a change is accepted, the agent lands the pull request.
Is symphony free to use? What license does it have?
symphony is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is symphony still maintained?
Updated within 6 months. The last commit to symphony was on Sep 15, 2026.
What can I use symphony for?
A team wants engineers to manage a backlog of implementation tasks rather than babysit coding agents. Experimenting with autonomous agent workflows in a trusted, non-production repository. Building an agent orchestrator in a different language from the provided spec.
How do I install symphony?
Read the Elixir reference implementation setup guide in elixir/README.md Or ask your coding agent to implement Symphony from the SPEC.md specification
Where can I get symphony?
The source code and installation instructions are on GitHub: https://github.com/openai/symphony

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects