SpecFlow

Open-source requirement & architecture engine for AI coding agents.

Quick start

npx specflow-cli init

Install Globally

npm i -g specflow-cli init

This will create a .specflow/ folder with project intelligence artifacts.

Commands

• specflow init - bootstrap a project and start requirement discovery.
• specflow status - show current state of the project.
• specflow generate - regenerate artifacts like architecture.md, modules.md, and JSON state files.
• specflow questions - answer templated questions (supports project/domain templates). These should be completed before using specflow ask.
• specflow ask - lets the AI ask follow-up questions based on the answered templates (AI will use the project summary + answered questions as context). The session ends automatically once the AI indicates it has enough information.
• specflow config - show and update your AI provider/model configuration.
• specflow doctor - validate the project structure and missing files.

Configuring AI Providers

SpecFlow can optionally use an AI provider to generate responses when you run specflow ask.

1. Run specflow init and select your provider + model.
2. Set the required API key in your environment or in an env file:
   • OpenAI: OPENAI_API_KEY
   • OpenRouter: OPENROUTER_API_KEY
   • Anthropic: ANTHROPIC_API_KEY
   • Ollama: (runs locally; no API key required)
   • Gemini: GOOGLE_API_KEY

SpecFlow will automatically load env variables from common env files (e.g. .env, .env.local, .specflow/.env, routes/.env, app/.env) if they exist. If the API key is not set, SpecFlow will still store your questions but won’t call the remote API.

Contributing

We welcome contributions! Please read CONTRIBUTING.md for guidance on how to contribute, run the project locally, and submit PRs.