OpenAI Open-Sources Symphony, a Coding Agent Orchestration Layer

Read in other languages: 한국어日本語
LLM May 2, 2026 By Insights AI 1 min read 1 views Source

OpenAI released Symphony as an open-source project on April 27, 2026. It is an orchestration layer designed to run coding agents continuously against an issue tracker, turning software project management boards into autonomous development pipelines.

How Symphony Works

Symphony is a long-running service that monitors an issue tracker — Linear in the reference implementation — and creates an isolated workspace for each open issue. A coding agent (Codex by default) runs inside each workspace, and Symphony handles crashes, restarts, CI monitoring, conflict resolution, rebasing, and shepherding pull requests toward review.

Results and Expansion

OpenAI teams that adopted Symphony saw a 500% increase in merged pull requests during the first three weeks of use. Though originally built for Codex, v1.1.0 introduced support for Claude Code, Gemini, and other agent runtimes via the Kata CLI — making Symphony a multi-model orchestration framework rather than an OpenAI-only tool.

The GitHub repository has attracted significant community contribution since launch, with developers extending support to additional issue trackers and agent backends.

Source: OpenAI Blog, GitHub

Share: Long

Related Articles

Comments (0)

No comments yet. Be the first to comment!

Leave a Comment