Quickstart

Get AI context for your Rails app in under 5 minutes.

Tools Reference · Recipes · Configuration · FAQ

Option A: In your Gemfile (recommended)

# 1. Add the gem
bundle add rails-ai-context --group development

# 2. Run the install generator
rails generate rails_ai_context:install

The generator asks two questions:

Which AI tools do you use? — Claude Code, Cursor, GitHub Copilot, OpenCode, Codex CLI, or all
Do you want MCP server support? — Yes (MCP mode) or No (CLI-only mode)

That’s it. Your AI tool now has live access to your schema, models, routes, controllers, views, and conventions.

[!TIP] The generator is idempotent — re-running it preserves existing config and only adds new sections.

Option B: Standalone (no Gemfile entry)

# 1. Install the gem globally
gem install rails-ai-context

# 2. Navigate to your Rails app
cd your-rails-app

# 3. Run interactive setup
rails-ai-context init

# 4. Start the MCP server
rails-ai-context serve

Works with every Ruby version manager (rbenv, rvm, asdf, mise, chruby, system). Learn more about standalone mode →

What just happened?

The install generator created:

File	Purpose
`.mcp.json` / `.cursor/mcp.json` / `.vscode/mcp.json` / `opencode.json` / `.codex/config.toml`	MCP auto-discovery — your AI tool detects these on project open
`CLAUDE.md` / `.cursor/rules/` / `.github/instructions/`	Static context rules your AI reads
`config/initializers/rails_ai_context.rb`	All configuration options
`.rails-ai-context.yml`	Config for standalone mode

Verify it works

# Check AI readiness (23 diagnostic checks)
rails ai:doctor          # In-Gemfile
rails-ai-context doctor  # Standalone

# Try a tool
rails 'ai:tool[schema]' table=users
rails 'ai:tool[model_details]' model=User
rails 'ai:tool[routes]' controller=users

See it work

[!IMPORTANT] These are CLI commands. When MCP is connected, your AI calls the same tools automatically — you never type these manually.

Three commands that show immediate value:

# 1. Trace a method across your entire codebase — one call
rails 'ai:tool[search_code]' pattern="your_method_name" match_type=trace

# 2. Full-stack feature analysis — models + controllers + routes + tests
rails 'ai:tool[analyze_feature]' feature=auth

# 3. Get your app's actual conventions — not generic Rails patterns
rails 'ai:tool[conventions]'

What to do next

Open your project in your AI tool — MCP auto-discovery kicks in
Ask your AI to describe your User model — it will call rails_get_model_details instead of guessing
Read Recipes for real-world workflows that show the tools in action

Quick reference

In-Gemfile	Standalone	What it does
`rails ai:serve`	`rails-ai-context serve`	Start MCP server
`rails 'ai:tool[NAME]'`	`rails-ai-context tool NAME`	Run any tool
`rails ai:tool`	`rails-ai-context tool --list`	List all tools
`rails ai:doctor`	`rails-ai-context doctor`	Diagnostics
`rails ai:watch`	`rails-ai-context watch`	Auto-regenerate on change
`rails ai:context`	`rails-ai-context context`	Regenerate context files

Next: Tools Reference →

Back to Home