From ticket to
merged PR.
Optio orchestrates AI coding agents across three modes: tasks that drive tickets to merged PRs, reusable agent workflows triggered on schedule or by webhook, and connections that give agents access to external tools — all automatically.
Up and running in minutes
$ git clone https://github.com/jonwiggins/optio.git
$ cd optio
$ ./scripts/setup-local.sh
# Dashboard at http://localhost:30310
# API at http://localhost:30400
Requires Docker Desktop with Kubernetes enabled. Full installation guide →
The complete task lifecycle
Every task flows through a seven-stage pipeline. Optio monitors each stage and automatically drives the task forward.
Intake
GitHub, GitLab, Linear, Jira, Notion, or manual
Queued
Enters the pipeline
Provisioning
Find or create pod
Running
Agent writes code
PR Opened
Opens pull request
CI & Review
Checks & feedback
Merged
Squash-merge & close
Intake
GitHub, GitLab, Linear, Jira, Notion, or manual
Queued
Enters the pipeline
Provisioning
Find or create pod
Running
Agent writes code
PR Opened
Opens pull request
CI & Review
Checks & feedback
Merged
Squash-merge & close
The feedback loop is
what makes it different.
Optio doesn't just run an agent and walk away. It watches the PR, feeds failures back to the agent, and keeps going until the work is done.
→ Task created from GitHub Issue #142
◇ Queued, waiting for pod...
⚡ Running claude-sonnet-4-6 in worktree
↗ PR #87 opened against main
✗ CI failed: lint errors in auth.ts
⚡ Resuming agent with CI context...
✓ CI passed, all checks green
◎ Review requested, awaiting approval
△ Review: "add error handling for edge case"
⚡ Resuming agent with review feedback...
✓ CI passed, review approved
✓ PR #87 squash-merged, Issue #142 closed
Everything you need to orchestrate AI agents
Built for teams that want to scale AI-assisted development without the manual overhead.
Autonomous Feedback Loop
CI fails? Agent resumes with failure context. Reviewer requests changes? Agent picks up the comments and pushes a fix. It keeps going until the PR merges.
Agent Workflows
Reusable, parameterized agent jobs that run standalone. Define prompt templates with variables, trigger manually, on a cron schedule, or via webhook. Auto-retry, cost tracking, and real-time log streaming included.
Connections
Give your agents access to external services at runtime. Built-in providers for Notion, Slack, Linear, GitHub, PostgreSQL, and Sentry, plus custom MCP servers and HTTP APIs with fine-grained access control.
Multi-Agent Support
Run Claude Code, OpenAI Codex, GitHub Copilot, Google Gemini, or OpenCode. Configure model, prompt template, and settings per repository. Launch review agents as subtasks with separate prompts.
Pod-per-Repo Isolation
One long-lived Kubernetes pod per repo with git worktree isolation. Multiple tasks run concurrently in separate worktrees. Multi-pod scaling and idle cleanup built in.
GitHub, GitLab, Linear, Jira & Notion Intake
Pull tasks from GitHub Issues, GitLab Issues, Linear, Jira, or Notion, or create them manually. One-click assign from the web UI kicks off the full pipeline.
Real-time Dashboard
Live log streaming, pipeline progress visualization, cost analytics, and cluster health monitoring. Watch your agents work in real time.
Self-Healing Pipeline
Auto-resume on CI failures, merge conflicts, and stale tasks. Auto-merge when CI passes and review is approved. Close linked issues on completion.
Built for production
Fastify API, Next.js dashboard, BullMQ workers, Drizzle on Postgres. Ships with a Helm chart for Kubernetes deployment.
Next.js · :3100
Fastify
Workers
Services
Repo Pod A
Repo Pod B
Workflow Pod
⚡ = Claude Code / Codex / Copilot / Gemini
Open source. Deploy on your infrastructure.
Optio is fully open source under the MIT license. Deploy on your own Kubernetes cluster with the Helm chart and start orchestrating AI agents in minutes.