What is this?
Symphony is an experimental system from OpenAI that turns project tasks into isolated, autonomous implementation runs, so teams manage the work itself instead of babysitting individual coding agents.
Symphony is an experimental system from OpenAI that turns project tasks into isolated, autonomous implementation runs, so teams manage the work itself instead of babysitting individual coding agents.How do I get it running?Follow the instructions in elixir/README.md to set up your environment and run the Elixir reference implementation, or ask your coding agent to set up Symphony for your repository based on the spec.
Follow the instructions in elixir/README.md to set up your environment and run the Elixir reference implementation, or ask your coding agent to set up Symphony for your repository based on the spec.Does my codebase need anything special before trying this?Symphony is designed to work best in codebases that have already adopted harness engineering, which gives agents a reliable way to run isolated tasks.
Symphony is designed to work best in codebases that have already adopted harness engineering, which gives agents a reliable way to run isolated tasks.Is this ready for production use?No, the README explicitly calls it a low-key engineering preview meant only for testing in trusted environments, not a finished product.
No, the README explicitly calls it a low-key engineering preview meant only for testing in trusted environments, not a finished product.How does Symphony know what work to pick up?In the demo, it monitors a Linear board for tasks and spawns agents to handle them automatically, without an engineer assigning each one manually.
In the demo, it monitors a Linear board for tasks and spawns agents to handle them automatically, without an engineer assigning each one manually.What proof does an agent give before its work is accepted?Agents report proof of work such as CI status, PR review feedback, complexity analysis, and a walkthrough video before a human accepts and lands the pull request.
Agents report proof of work such as CI status, PR review feedback, complexity analysis, and a walkthrough video before a human accepts and lands the pull request.Why use Symphony instead of running Codex directly?It shifts engineers from supervising individual coding-agent sessions to managing work at a higher level, letting agents land PRs safely once their work is accepted.
It shifts engineers from supervising individual coding-agent sessions to managing work at a higher level, letting agents land PRs safely once their work is accepted.Which topics is it in?elixir, project-management, harness-engineering, autonomous-agents, github-actions, ci, orchestration, agent-orchestration, multi-agent, agents, coding-agent, ai-agent
elixir, project-management, harness-engineering, autonomous-agents, github-actions, ci, orchestration, agent-orchestration, multi-agent, agents, coding-agent, ai-agent