# Optio

> Workflow orchestration for AI coding agents that turns tasks from GitHub Issues or Linear tickets into merged pull requests with autonomous CI feedback loops.

Optio is an open-source workflow orchestration platform that automates the entire lifecycle of AI-powered code changes, from task intake to merged pull request. It accepts work items from GitHub Issues, Linear tickets, or manual task creation through its web dashboard, then provisions isolated Kubernetes environments where AI coding agents like Claude Code or OpenAI Codex execute the work.

The core differentiator is its autonomous feedback loop. When CI fails after the agent opens a pull request, Optio automatically resumes the agent with the failure context so it can fix the build. When a code reviewer requests changes, the agent receives the review feedback and pushes corrections. When merge conflicts arise, the agent rebases and resolves them. Only when CI passes and reviews are approved does Optio squash-merge the PR and close the linked issue.

Optio uses a pod-per-repo architecture on Kubernetes, where each repository gets a long-lived pod with git worktree isolation for concurrent tasks. This enables multiple agents to work on different branches of the same repository simultaneously without interference. The system includes multi-pod scaling and automatic idle cleanup to manage cluster resources efficiently.

The platform ships with a Next.js dashboard providing real-time visibility into running agents, live log streaming, pipeline progress tracking, cost analytics, and cluster health monitoring. Per-repository configuration allows teams to customize the AI model, prompt templates, container images, concurrency limits, and setup commands for each project.

Optio includes a built-in code review agent that launches as a subtask with its own prompt and model configuration, enabling automated review before human reviewers are notified. The system supports multi-provider OAuth authentication through GitHub, Google, and GitLab.

The tech stack is built on Fastify for the API server, BullMQ with Redis for job queuing and pub/sub messaging, Drizzle ORM with PostgreSQL for persistence, and Turborepo with pnpm for monorepo management. Production deployment is supported through a Helm chart for Kubernetes clusters, while local development uses Docker Desktop with Kubernetes enabled.

Optio provides container images for multiple language ecosystems including Node.js, Python, Go, and Rust, along with a full image that combines all runtimes. The project includes comprehensive setup scripts that bootstrap the local infrastructure including database migrations and environment configuration.

## Features
- Autonomous feedback loop that resumes AI agents on CI failures, merge conflicts, and review feedback
- Pod-per-repo Kubernetes architecture with git worktree isolation for concurrent task execution
- Built-in code review agent launched as a subtask with separate prompt and model configuration
- Per-repository configuration for model selection, prompt templates, container images, and concurrency limits
- Task intake from GitHub Issues, Linear tickets, or manual creation through the web dashboard
- Real-time Next.js dashboard with live log streaming, pipeline progress, and cost analytics
- Automatic squash-merge and issue closure when CI passes and reviews are approved
- Multi-provider OAuth authentication supporting GitHub, Google, and GitLab
- Helm chart for production Kubernetes deployment with configurable resource limits
- Container images for Node.js, Python, Go, Rust, and combined full-stack environments
- PR watcher that polls every 30 seconds for CI status, review state, and merge readiness
- Subtask system enabling hierarchical agent task decomposition

## Integrations
Claude Code, OpenAI Codex, GitHub, GitHub Issues, GitHub Actions, Linear, Kubernetes, Docker, PostgreSQL, Redis, BullMQ, Helm

## Platforms
LINUX, MACOS, WEB

## Pricing
Open Source

## Version
0.1.0

## Links
- Website: https://github.com/jonwiggins/optio
- Documentation: https://github.com/jonwiggins/optio/blob/main/README.md
- Repository: https://github.com/jonwiggins/optio
- EveryDev.ai: https://www.everydev.ai/tools/optio