AI Commands

Fastpy includes AI-powered resource generation and configuration generators for popular AI coding assistants.

ai:config

Configure your AI provider for code generation.

# Show current configuration and available providers
fastpy ai:config

# Set AI provider
fastpy ai:config -p anthropic
fastpy ai:config -p openai
fastpy ai:config -p ollama

# Test your connection
fastpy ai:config --test

Supported Providers

Provider	Model	API Key Variable
Anthropic	Claude Sonnet	ANTHROPIC_API_KEY
OpenAI	GPT-4o	OPENAI_API_KEY
Ollama	Llama 3.2 (local)	None required

Quick Setup

# 1. Set provider
fastpy ai:config -p anthropic

# 2. Set API key (add to ~/.bashrc or ~/.zshrc)
export ANTHROPIC_API_KEY=your-key-here

# 3. Test connection
fastpy ai:config --test

Getting API Keys

• Anthropic (Claude): console.anthropic.com
• OpenAI (GPT-4): platform.openai.com
• Ollama: Free, runs locally - ollama.ai

ai

Generate complete resources using natural language descriptions.

fastpy ai "Create a blog with posts, categories, and tags"

Usage

# Basic usage - generates commands and asks for confirmation
fastpy ai "E-commerce with products, orders, and customers"

# Auto-execute generated commands
fastpy ai "User management system" --execute

# Preview commands without executing
fastpy ai "Blog system" --dry-run

# Use a specific provider
fastpy ai "Task manager with projects" --provider ollama

Examples

# Blog system
fastpy ai "Create a blog with posts, categories, tags, and comments"

# E-commerce
fastpy ai "E-commerce store with products, categories, cart, and orders"

# Project management
fastpy ai "Project management with projects, tasks, and team members"

# Social media
fastpy ai "Social network with users, posts, likes, and followers"

Check Configuration

View your current AI configuration:

fastpy config

Output:

AI Configuration

  Provider: anthropic
  ANTHROPIC_API_KEY: Set

Set provider with: export FASTPY_AI_PROVIDER=anthropic|openai|ollama

Using Ollama (Free, Local)

For a free, local option, use Ollama:

# Install Ollama
curl -fsSL https://ollama.ai/install.sh | sh

# Pull a model
ollama pull llama3.2

# Set as provider
export FASTPY_AI_PROVIDER=ollama

# Use it
fastpy ai "Create a todo app with tasks and categories"

Configure Ollama:

export OLLAMA_MODEL=llama3.2      # Model to use
export OLLAMA_HOST=http://localhost:11434  # Ollama server URL

---

AI Assistant Configuration

Fastpy also includes configuration generators for popular AI coding assistants. These configuration files help AI assistants understand your project structure, conventions, and available commands.

Configuration

Generate AI assistant configurations using the CLI:

# Interactive mode - choose from menu
fastpy init:ai

# Direct selection
fastpy init:ai claude
fastpy init:ai copilot
fastpy init:ai cursor
fastpy init:ai gemini

Supported AI Assistants

Claude Code

File: CLAUDE.md

For use with Claude Code - Anthropic's CLI tool.

fastpy init:ai claude

GitHub Copilot

File: .github/copilot-instructions.md

For use with GitHub Copilot in VS Code, JetBrains, and other IDEs.

fastpy init:ai copilot

Cursor AI

File: .cursorrules

For use with the Cursor AI-powered code editor.

fastpy init:ai cursor

Google Gemini

File: .gemini/instructions.md

For use with Google Gemini Code Assist.

fastpy init:ai gemini

What's Included in the Configuration

The generated configuration file includes comprehensive project context:

Development Commands

# Activate virtual environment
source venv/bin/activate

# Start development server
fastpy serve
uvicorn main:app --reload

Code Generation Commands

# Generate complete resource
fastpy make:resource Post -f title:string:required -m -p

# Individual generators
fastpy make:model Post -f title:string:required -m
fastpy make:controller Post
fastpy make:route Post --protected
fastpy make:service Post
fastpy make:repository Post
fastpy make:middleware Logging
fastpy make:test Post
fastpy make:factory Post
fastpy make:seeder Post
fastpy make:enum Status -v active -v inactive
fastpy make:exception NotFound -s 404

Database Commands

fastpy db:migrate                    # Run pending migrations
fastpy db:migrate -m "Add posts"     # Generate + run migrations
fastpy db:make "Add slug"            # Generate migration only
fastpy db:rollback
fastpy db:fresh
fastpy db:seed

Project Architecture

• Directory structure and purpose of each folder
• Naming conventions (tables, models, files)
• Base model features (soft deletes, timestamps)

API Documentation

• Authentication endpoints
• Standard response formats
• Custom exceptions
• Middleware details
• Health check endpoints

Field Types and Validation

• All available field types
• Validation rules and syntax
• Examples for common patterns

Customizing the Configuration

After generating, you can customize the configuration file to include:

• Project-specific instructions
• Custom conventions
• Business logic guidelines
• Team coding standards

Example customization at the end of the file:

## Project-Specific Instructions

- Use snake_case for all API endpoints
- All monetary values use the `money` field type
- Include pagination on all list endpoints
- Always add soft delete support to models

Benefits of AI Configuration

1. Consistent Code Generation - AI follows your project patterns
2. Faster Development - AI understands available commands
3. Reduced Errors - AI knows correct syntax and conventions
4. Better Suggestions - AI provides contextually relevant code

Example AI Interactions

With Claude Code

You: Create a Product model with name, price, and category

Claude: I'll create a Product model using Fastpy conventions:

fastpy make:resource Product \
  -f name:string:required,max:200 \
  -f price:money:required,ge:0 \
  -f category_id:integer:required,foreign:categories.id \
  -m -p

With GitHub Copilot

When you type a comment, Copilot suggests code following Fastpy patterns:

# Create endpoint to get products by category
@router.get("/category/{category_id}")
async def get_by_category(
    category_id: int,
    session: AsyncSession = Depends(get_session)
):
    # Copilot suggests the correct controller method
    return await ProductController.get_by_category(session, category_id)