Spec-driven development for AI teams

Coordinate Multiple AI Agents. Ship Better Code.

Aigon turns multi-agent coding into a deliberate system. Research in parallel, implement with worktree arenas, review alternatives, and keep every decision linked back to a spec.

See the workflow Read docs

Git-native workflow with branch and worktree isolation
Agent-agnostic: Claude, Gemini, Cursor, Codex, and more
No telemetry, no lock-in, MIT licensed

Terminal demo

$ aigon research-setup 05 cc gg cx
$ aigon feature-setup 01 cc gg cx cu
$ aigon feature-eval 01
$ aigon feature-done 01 cx

Research → Features → Code Review → Feedback → Repeat

The problem

AI agents multiply output fast, but coordination breaks down even faster.

Terminal tabs everywhere

Parallel agents generate momentum, but tracking what each one did and why becomes manual overhead.

Decisions get lost

Without spec-linked workflows, merge decisions and implementation tradeoffs disappear into chat history.

No clean comparison step

Teams rarely compare multiple implementations side-by-side before merging the best approach.

Feedback loops stay disconnected

User feedback, research, and implementation often live in separate tools with no shared lifecycle.

Core features

Everything needed to run a disciplined multi-agent engineering loop.

Research Arena

Run parallel research before coding

Spin up several agents on one topic and compare findings before writing a single line of product code.

aigon research-setup 05 cc gg cx

Feature Workflow

Spec to implementation with traceability

Move features through clear folders and logs so project state is visible from the file tree.

aigon feature-implement 01

Arena Mode

Compare multiple implementations

Create separate worktrees for each agent, then evaluate alternatives and merge the winner.

aigon feature-setup 01 cc gg cx cu

Feedback Loop

Capture and triage real user input

Convert incoming feedback into structured tasks and link outcomes back to specs and decisions.

aigon feedback-triage 14

The big picture

Aigon keeps research, delivery, and feedback in one continuous system.

Research

Explore alternatives, constraints, and market context with multiple agents.
Features

Convert winning ideas into specs that define scope, intent, and success criteria.
Code Review

Evaluate implementations, merge the strongest option, and keep rationale in logs.
Feedback

Capture user signals and route them back into research and planning.

How it works

From idea to production in five explicit steps.

Initialize

aigon init
aigon install-agent cc gg cx

Research

aigon research-create "auth-strategy"
aigon research-setup 05 cc gg

Implement

aigon feature-create jwt-auth
aigon feature-now jwt-auth

Capture feedback

aigon feedback-create "Login is slow"

Triage

aigon feedback-triage 15

Documentation

Quick start fast, then go deep with reference docs.

git clone https://github.com/jayvee/aigon-site.git
cd aigon-site
aigon init
aigon feature-create landing-page-improvements

Workflow Guide

Research, specs, implementation, evaluation, and completion flow.

Project README

Setup, deployment on Cloudflare Pages, and local usage.

Tech & philosophy

Open source, git-native, and intentionally simple.

Aigon is built for teams who want disciplined AI-assisted engineering, not opaque automation.

Open SourceMIT licensed, no paid-tier lockouts.
No TelemetryYour code and prompts stay in your workflow.
Agent-AgnosticWorks with whichever coding agents your team chooses.
Zero Build OverheadThis site ships as plain HTML and CSS.

Community

Help shape the next generation of collaborative AI development.

Contribute specs, improve workflows, and share real-world patterns for running multi-agent engineering teams effectively.

Star on GitHub Contributing guide