Generate AI Context Files for Any Codebase in 60 Seconds

FAF Wizard generates a project.faf file - a machine-readable, IANA-registered YAML format (application/vnd.faf+yaml) - that gives AI tools persistent context about how to help build a project, as distinct from README/docs, which explain how humans use it. It runs a three-phase, roughly 60-second pipeline: Stack Detection, scanning package.json, Cargo.toml, pyproject.toml, and directory structure to identify frameworks, deployment targets, and testing setup; Context Mining, pulling a project description from the README, architecture patterns from code structure, and dependency info; and Optimization, generating a focused 33-slot project.faf, validating it against the format spec, and scoring AI-readiness with concrete suggestions to reach the next tier, for example "add API documentation (+5%)." The vendor reports 99.2% detection accuracy across 153+ formats and an average 12.3-second generation time across an analysis of 8,400+ projects, with 94% of projects reaching Bronze tier or higher; per-project-type figures in its own success-metrics table range from 76% average score for legacy PHP to 91% for Python Django. It supports migrating existing AI context files (faf migrate --from .cursorrules, CLAUDE.md, or README.md) and syncing one .faf file to every AI tool's own format (faf sync --target all) - Claude Code reads .faf natively, Cursor syncs to .cursorrules, Gemini CLI converts to GEMINI.md, Windsurf syncs to .windsurfrules, and ChatGPT can read the plain YAML.

Use faf-wizard for quick, fully automatic project onboarding - "just make it work" in a time-constrained scenario, on a new project, a legacy codebase, a famous OSS repo, a side project, a client handoff, or a team project. Use the related faf-expert skill instead when you need fine-tuned championship-level scoring of 95% or higher, complex MCP server configuration, multi-platform sync management, or enterprise deployment patterns - faf-wizard optimizes for speed and automation, not maximum score.

Input: any project directory or a GitHub repository URL, via the browser extension. Output: a project.faf YAML file with project metadata, detected stack, and human context, plus an AI-readiness score and tier such as Bronze or Silver, with specific, percentage-quantified suggestions for improving it.

• CLI (npm install -g faf-cli, then faf auto), an MCP server (claude-faf-mcp) for Claude Code, and a Chrome browser extension
• AI tools: Claude Code, Cursor, Gemini CLI, Windsurf, and any tool that can read plain YAML
• Migration from existing context files: .cursorrules, CLAUDE.md, README.md
• MIT licensed, with no credentials ever stored in .faf files

Developers who want any codebase - new, legacy, open source, or a personal side project - to have persistent, portable AI context in about a minute, without hand-writing or maintaining a context file for every AI tool separately.