Forge is a disposable meta-tool for AI coding agents. One command audits your repo - stack, layers, conventions - and generates architecture guardrails, quality patterns, and test traceability. Then you can remove Forge. Artifacts work standalone. Unlike static rule files (Cursor Rules, CLAUDE.md) - a 7-gate audit with 50+ patterns that prevent AI shortcuts. Unlike Devin - you keep control. Full pipeline from design to tested code. Works with Claude Code, Cursor, any AI agent. From EUR 29.
No reviews yetBe the first to leave a review for Forge DevKit
Fin Startups get Fin free for a year + 93% off Intercom
Promoted
Maker
📌
I've been a backend engineer for 10+ years. When AI coding agents became my daily workflow, I noticed a pattern: every session started with me re-explaining my architecture, conventions, and quality standards. The AI would still ignore half of it.
I tried writing detailed CLAUDE.md files manually. It helped, but took hours and drifted as the codebase evolved. Worse - there was no connection between those rules and actual test generation.
So I built a system that does it automatically. One command scans your repo - stack, layers, patterns - and generates everything: architecture guardrails, quality gates, test traceability. The key insight was making it disposable: Forge generates artifacts, then you can uninstall it. No lock-in, no dependency.
It started as a single plugin for architecture audit. Now it's 15 modules covering the full pipeline - from feature design to parallel AI agents working on your backlog without merge conflicts.
Works with Claude Code, Cursor, and any AI agent that reads project files.
The approach evolved from "better rules" to "full dev pipeline where AI respects your project at every step." Happy to answer anything about the architecture, approach, or pricing decisions.