One spec, many agents, the best diff wins.

Aigon orchestrates Claude Code, Gemini CLI, Codex, and Cursor from a single spec. Competing implementations, cross-agent review, merge the survivor.

Open source · MIT · Bring your own subscriptions

You can run Aigon in four modes — Drive, Fleet, Autopilot, and Swarm.

1. Kick off the dark-mode feature in Fleet mode, with Cursor and Claude Code implementing in isolated worktrees

Fleet Mode — implement the dark-mode feature with Cursor and Claude Code in isolated worktrees

2. Both agents (Cursor, Claude Code) have completed; review each worktree's running site, then choose an evaluation agent/model (Kimi K2.6)

Both agents have completed; review each worktree's running site, then choose an evaluation agent/model

3. Evaluation is complete — merge the winner (Claude Code), include the best parts from the Cursor implementation

Evaluation complete — winner selected and merged

Why Aigon

Different models catch different bugs.

Run implementations across model families. Review with a different agent than the one that wrote the code.

Your subscriptions, zero markup.

Aigon orchestrates your agent CLIs directly. No billing, no metering, no third-party systems.

Local data, simple tools.

Git worktrees, tmux sessions, Markdown specs. Attach to any session, control any agent, any time.

CLI in action

Aigon slash commands can be run from any agent.

How it works

Research, build, learn — on repeat.

Aigon modes

Four ways to run, same spec.

One agent, hands on.

You guide a single agent through the spec. Full control, full visibility. The agent works in a dedicated branch, you review and submit when ready.

$ aigon feature-start 07
$ aigon feature-do 07
$ aigon feature-close 07

The steps

From idea to merged code.

Create a feature spec and move it into the backlog.

Your workflow, visualised

A Kanban board for spec-driven development.

The Aigon Dashboard is a visual way to run and monitor your spec-driven workflow. When you have multiple concurrent features in active development across many agents and tmux sessions, the dashboard provides a single pane to control and view progress. Drag features across columns, launch agent sessions with one click, and watch your pipeline move in real time. A web-based tmux session view allows you to interact with your agents without ever leaving the dashboard.

Remote Access

Monitor agents from your phone

The dashboard runs on your local network — open it on your phone or tablet over LAN, or use Tailscale to check on your agents from anywhere. Monitor from the couch.

Aigon Dashboard viewed on iPhone Safari showing active agent sessions across multiple repositories Peeking into a live Claude agent session from iPhone — viewing real-time tool calls and reasoning

Workflow Intelligence

Know if your AI workflow is actually working

Aigon comes with charts that track the metrics that help you answer questions such as: Are features shipping faster? Is rework going down? Is token usage increasing per feature?

Aigon Reports — five aligned charts showing features completed, commits, cycle time, commits per feature, and rework ratio

Feature velocity

Features completed and commits over time — stacked to show feature-attributed vs non-feature work.

Cycle time & focus

Median cycle time and commits-per-feature trending over time. See whether features are getting tighter.

Rework ratio

Percentage of fix commits per period. Trending down means agents are landing correct code on the first pass.

Aigon Pro

From running features to operating a workflow.

Aigon Pro adds advanced features and insights: autonomous runs on a schedule, agent quality metrics with AI coaching, benchmarks, and portable state across machines.

  • First-pass rate, rework ratio & agent leaderboard
  • Five synchronized trend charts — cycle time, commits, rework
  • Token usage & cost attributed per agent and phase
  • AI-generated observations & coaching
  • Agent benchmarks — reproducible perf timings across CC, Codex, Gemini, OpenRouter, Kimi
  • Reusable autonomous workflows — named stages saved to git, launched from CLI or dashboard
  • Scheduled autonomous runs — overnight, or aligned with quota refresh
  • Recurring features — cadence-driven template batches (weekly, quarterly)
  • Set-level autonomous orchestration — start, stop, resume entire feature sets
  • Aigon Sync — private Git vault for portable workflow state across machines
Explore Aigon Pro
Aigon Pro — AI-powered insights and coaching dashboard