OpenAI releases Symphony, an open-source spec that lets AI agents manage their own tasks from Linear project trackers instead of requiring constant human supervision.

OpenAI released Symphony, an open-source spec with a reference implementation that connects task trackers like Linear to Codex agents. Some internal teams saw merged pull requests jump sixfold in the first three weeks, according to OpenAI.

Symphony uses Linear as a state machine (a system that tracks work status across defined stages). Each open ticket gets its own Codex agent and dedicated workspace that runs until the task is done. Agents pull work directly from the tracker instead of developers juggling multiple sessions, and can file new tickets if they spot issues like performance problems or refactoring opportunities.

The spec is shipped as a Markdown file that agents can implement themselves. The core workflow—accept ticket, check out repo, set status, attach PR, attach video—lives in a WORKFLOW.md file that Symphony hands to agents as a guide. OpenAI does not plan to maintain Symphony as a standalone product; the community has already shipped forks, including one for Anthropic's Claude Code with GitHub Issues. Code and specs are available on GitHub.