#2121st of 29 in Coding agents and assistants
symphony
Spec and Elixir reference for agents that pick up Linear tasks and open PRs
Live demo ↗
(opens in a new tab)Documentation ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 27.7k
- License
- Apache-2.0
- Last commit
- Sep 2026
- Last release
- Sep 2026
- Language
- Elixir
Checked on GitHub.
OpenAI is hiring: 808 open roles
Overview
Symphony watches a Linear board for work and spawns isolated agent runs for each task. Agents report proof of work such as CI status, PR review feedback, complexity analysis and walkthrough videos, and land the PR once accepted. The repo ships a language-agnostic SPEC.md and an experimental Elixir reference implementation.
Who it is for: Teams wanting to delegate tracked tickets to coding agents
Strengths
- Language-agnostic SPEC.md lets you implement it in any language
- Each task runs as an isolated, autonomous implementation run
- Agents attach proof of work: CI status, review feedback, walkthrough videos
- Apache-2.0 licensed reference implementation in Elixir
Weaknesses
- Described as a low-key engineering preview for trusted environments only
- Reference implementation is experimental
- Expects codebases that have already adopted harness engineering
- RAM, ports and model support are not stated in the main README
What it needs
- Needs Linear, Codex
- Models: Codex
Also in Coding agents and assistants
See all 29| Rank | Project | Score |
|---|---|---|
| 1 | claw-codeRust CLI coding agent harness using Anthropic and OpenAI-compatible APIs | 83 out of 100 |
| 2 | opencodeTerminal coding agent with build and plan modes, plus a desktop app | 79 out of 100 |
| 3 | orcaDesktop app that runs CLI coding agents in parallel git worktrees | 77 out of 100 |
| #4 | oh-my-openagentTerminal coding agent that fans work out across many models | 77 out of 100 |
| #5 | openinterpreterTerminal coding agent, a Codex fork tuned for low-cost models | 76 out of 100 |