Govern AI Code Your Way
CLI for total control. Drop-in for your existing tools.
Website •
Documentation •
Download •
Changelog
ArchonDev is an AI-powered development governance system that prevents your AI coding assistant from making architectural mistakes, forgetting context, and introducing regressions.
The Problem: AI coding assistants are powerful but unreliable. They hallucinate, ignore your architecture, forget what they learned, and break things they just fixed.
The Solution: ArchonDev adds governance — architectural rules your AI must follow, quality gates before every change, and persistent learning across sessions.
AI terminal that governs itself. Total control with architecture enforcement, dependency tracking, and security sentinels.
npm install -g archondevThen run:
archonGovernance files for any IDE or AI tool. Drop into Cursor, Claude Code, Windsurf, or any AI coding assistant.
Download for your IDE:
| IDE | Download |
|---|---|
| Cursor | cursor-package.zip (~20KB) |
| Claude Code / Amp | claude-amp-package.zip (~20KB) |
| Google Gemini | gemini-package.zip (~20KB) |
| Windsurf / Codeium | windsurf-package.zip (~20KB) |
| VS Code + Copilot | vscode-copilot-package.zip (~20KB) |
| OpenAI Codex | codex-package.zip (~20KB) |
| Generic (any AI) | generic-package.zip (~20KB) |
What's in the package:
| File | Purpose |
|---|---|
ARCHITECTURE.md |
Governance constitution — components, invariants, protected paths, accessibility config |
AGENTS.md |
AI instructions — code review, task extraction, accessibility, SEO, GEO, memory management |
DEPENDENCIES.md |
Regression prevention — tracks what breaks when you change files |
progress.txt |
Learning log — persists knowledge across sessions |
.archon/config.yaml |
Configuration file |
archondev-scenarios/ |
Smart onboarding for new/existing/continuing projects |
examples/ |
Sample workflows and patterns |
| IDE-specific rules | .cursorrules, CLAUDE.md, GEMINI.md, etc. |
Usage:
- Download the package for your IDE
- Unzip to your project root
- Tell your AI:
"read ARCHITECTURE.md"
Real-time violations + quick-fix suggestions. Press Ctrl+. (Cmd+.) for lightbulb menu with auto-fixes.
- 📐 Architectural Governance — Define components, boundaries, and invariants your AI must respect
- ⚡ Quality Level / Posture — Right-sized architecture (prototype/production/enterprise)
- 🔗 Dependency Tracking — Know what breaks before you change it (
DEPENDENCIES.md) - 🧠 Learning Persistence — AI remembers patterns across sessions via
progress.txt - 🛡️ Quality Gates — Every change must pass before commit
- ⚡ Quality Level / Posture — Tell AI how rigorous to be
prototype: Fast iteration, minimal governance, skip complex patternsproduction: Secure defaults, basic monitoring, modular design (default)enterprise: Full governance with audit logging, RBAC, SLOs, compliance- Applies to architecture, code generation, AND code review
- Prevents over-engineering for simple projects
- Trigger: Set
qualityLevel.posturein ARCHITECTURE.md
- 🔍 SEO Optimization — Automated meta tags, Open Graph, Twitter Cards
- AI scans, identifies gaps, generates missing tags
- User approves before changes are applied
- Trigger:
seo check,seo fix,add open graph
- 🤖 GEO for AI Search — Optimize for ChatGPT, Perplexity, Claude citations
- AI generates 3 candidate 7-word brand phrases
- AI generates 3 candidate 50-word descriptions
- User selects preferred identity
- JSON-LD schemas for AI comprehension
- Trigger:
geo identity,geo schema
- ♿ Pre-Deploy Accessibility Check — WCAG 2.2 AA compliance before going live
- Legal liability warnings (ADA, EAA, Section 508)
- Auto-fix for common issues (contrast, alt text, focus)
- WCAG 2.2 AA badge for compliant sites
- CLI:
archon a11y check,archon a11y fix,archon a11y badge
- 📋 Task Extraction Protocol — AI confirms all items in multi-item requests before starting
- Prevents lost requirements (AI often forgets items 3+ in a list)
- Trigger phrases:
plan these tasks,task status,what's on my list
- 🚀 Smart Onboarding — Detects new project, existing project, or continuing session
- 🔍 Code Review Mode — AI reviews code without modifying it
- 📦 Local Database — Optional SQLite for tracking atoms and learnings
- 💡 VS Code Extension — Real-time diagnostics with quick-fix suggestions
| Command | Description |
|---|---|
archon |
Interactive mode |
archon plan <description> |
Create governed work item |
archon execute <atom-id> |
Execute with quality gates |
archon review init |
Initialize code review |
archon a11y check |
Run WCAG 2.2 AA audit |
archon a11y fix |
Auto-fix accessibility issues |
archon deps list |
View dependency rules |
archon watch |
Live TUI dashboard |
archon seo check |
Run SEO meta tag audit |
archon seo fix |
Apply recommended SEO fixes |
archon geo identity |
Generate brand identity phrases |
archon geo schema |
Generate JSON-LD schemas |
| Tier | Cost | What You Get |
|---|---|---|
| Free | $0 | Ultra-cheap models (GPT-5-nano, Gemini Flash-Lite) |
| Credits | Pay as you go | All models, 10% service fee |
| BYOK | $0 | Use your own API keys |
No subscriptions. No commitments.
- Website: archondev.io
- Documentation: archondev.io/docs
- Changelog: archondev.io/changelog
- Download Lite Packages: archondev.io/download
MIT © Jumping Ahead Corp.
Built with the Ralph Protocol — fresh context, persistent learning.
