Skip to main content

AI Features

ReqSmith integrates AI capabilities to enhance your API testing experience with intelligent suggestions, request optimization, and automated analysis. Currently powered by Google's Gemini AI, these features help you work more efficiently and catch potential issues early.

Overview​

ReqSmith's AI features include:

  • Request suggestions: AI-powered recommendations for API endpoints
  • Response analysis: Intelligent analysis of API responses
  • Error diagnosis: AI assistance with debugging failed requests
  • Template optimization: Suggestions for improving request templates
  • Pattern recognition: Identify common API usage patterns
  • Documentation generation: Auto-generate API documentation from requests

Setup and Configuration​

Enable AI Features​

AI features are optional and require configuration:

# Check if AI is available
reqsmith ai status

# Configure Gemini API key
reqsmith config set ai.gemini_api_key "your-api-key-here"

# Enable AI features
reqsmith config set ai.enabled true

# Set AI provider (currently only Gemini supported)
reqsmith config set ai.provider "gemini"

API Key Setup​

Get your Gemini API key:

  1. Visit Google AI Studio
  2. Create a new API key
  3. Configure ReqSmith:
# Set API key directly
reqsmith config set ai.gemini_api_key "your-api-key"

# Or use environment variable
export GEMINI_API_KEY="your-api-key"

# Or store in .env file
echo "GEMINI_API_KEY=your-api-key" >> .env

AI Configuration Options​

Customize AI behavior:

# Set AI model (default: gemini-pro)
reqsmith config set ai.model "gemini-pro"

# Set creativity level (0.0-1.0, default: 0.3)
reqsmith config set ai.temperature 0.3

# Set response length preference
reqsmith config set ai.max_output_tokens 1024

# Enable/disable specific AI features
reqsmith config set ai.suggestions_enabled true
reqsmith config set ai.analysis_enabled true
reqsmith config set ai.documentation_enabled true

Request Suggestions​

Getting Suggestions​

AI can suggest API endpoints and request improvements:

# Get suggestions for a URL pattern
reqsmith ai suggest "https://api.github.com"

# Get suggestions for specific functionality
reqsmith ai suggest --describe "get user repositories"

# Get suggestions for a specific API
reqsmith ai suggest --api "GitHub REST API" --action "list issues"

Contextual Suggestions​

AI provides context-aware suggestions:

# Suggest based on current environment
reqsmith ai suggest --env production --describe "health check"

# Suggest based on previous requests
reqsmith ai suggest --from-history --pattern "user management"

# Suggest improvements for existing template
reqsmith ai suggest --template github-user --improve

Example Suggestions​

$ reqsmith ai suggest "https://api.github.com"

🤖 AI Suggestions for GitHub API:

1. GET /user - Get authenticated user
2. GET /users/{username} - Get specific user
3. GET /user/repos - List user repositories
4. GET /repos/{owner}/{repo} - Get repository details
5. GET /repos/{owner}/{repo}/issues - List repository issues

Would you like me to create templates for any of these endpoints?

Response Analysis​

Automatic Analysis​

AI automatically analyzes API responses:

# Make request with AI analysis
reqsmith get https://api.github.com/users/octocat --analyze

# Analyze specific response aspects
reqsmith get https://api.example.com/data --analyze-performance
reqsmith get https://api.example.com/data --analyze-structure
reqsmith get https://api.example.com/data --analyze-security

Analysis Types​

Different types of AI analysis:

Performance Analysis​

# Analyze response time and size
reqsmith ai analyze-performance --request-id 12345

# Example output:
# ⚡ Performance Analysis:
# - Response time: 245ms (Good)
# - Payload size: 2.3KB (Optimal)
# - Recommendation: Consider caching for repeated requests

Structure Analysis​

# Analyze response structure and schema
reqsmith ai analyze-structure --request-id 12345

# Example output:
# 📋 Structure Analysis:
# - Response format: JSON
# - Schema: Well-formed with consistent field types
# - Pagination: Uses 'next' and 'previous' links
# - Recommendation: Response follows REST best practices

Security Analysis​

# Analyze security aspects
reqsmith ai analyze-security --request-id 12345

# Example output:
# 🔒 Security Analysis:
# - Headers: Missing security headers (X-Frame-Options)
# - Authentication: Bearer token properly implemented
# - Data exposure: No sensitive data in response
# - Recommendation: Add security headers for production

Error Diagnosis​

AI-Powered Debugging​

Get AI assistance for failed requests:

# Analyze last failed request
reqsmith ai diagnose --last-error

# Diagnose specific error
reqsmith ai diagnose --request-id 12345

# Diagnose with additional context
reqsmith ai diagnose --request-id 12345 --include-logs

Error Analysis Examples​

$ reqsmith ai diagnose --last-error

🔍 Error Diagnosis:

Request: POST https://api.example.com/users
Status: 422 Unprocessable Entity
Error: Validation failed

🤖 AI Analysis:
The request failed due to validation errors. Common causes:
1. Missing required fields in request body
2. Invalid data format (expected email format)
3. Duplicate username constraint violation

💡 Suggested fixes:
1. Check required fields: name, email, username
2. Validate email format before sending
3. Use unique username or check availability first

Would you like me to create a corrected template?

Template Optimization​

AI Template Improvements​

Get suggestions for improving templates:

# Analyze template for improvements
reqsmith ai optimize-template github-user

# Get specific optimization suggestions
reqsmith ai optimize-template github-user --focus error-handling
reqsmith ai optimize-template github-user --focus performance
reqsmith ai optimize-template github-user --focus security

Template Generation​

AI can generate templates from descriptions:

# Generate template from description
reqsmith ai create-template \
  --name "user-crud" \
  --description "Complete CRUD operations for user management"

# Generate from example API documentation
reqsmith ai create-template \
  --from-docs "https://docs.example.com/api" \
  --endpoints "users"

Template Analysis​

Analyze existing templates:

# Analyze template completeness
reqsmith ai analyze-template user-management

# Check template best practices
reqsmith ai check-template user-management --best-practices

# Suggest template variables
reqsmith ai suggest-variables user-management

Pattern Recognition​

Usage Pattern Analysis​

AI identifies patterns in your API usage:

# Analyze request patterns
reqsmith ai analyze-patterns

# Find common workflows
reqsmith ai find-workflows --days 30

# Identify optimization opportunities
reqsmith ai suggest-optimizations --history

Pattern Examples​

$ reqsmith ai analyze-patterns

📊 API Usage Patterns Detected:

1. Authentication Pattern:
   - Always GET /auth/token before other requests
   - Token expires every 1 hour
   - Suggestion: Implement automatic token refresh

2. Data Fetching Pattern:
   - GET /users/{id} followed by GET /users/{id}/profile
   - Could be optimized with single endpoint
   - Suggestion: Use expanded response (?expand=profile)

3. Error Handling Pattern:
   - 401 errors always followed by re-authentication
   - Suggestion: Implement automatic retry with auth refresh

Documentation Generation​

Auto-Generate Documentation​

AI can create documentation from your requests:

# Generate docs for all templates
reqsmith ai generate-docs

# Generate docs for specific template
reqsmith ai generate-docs --template github-api

# Generate API specification
reqsmith ai generate-spec --format openapi

# Generate usage examples
reqsmith ai generate-examples --template user-management

Documentation Types​

Different documentation formats:

# Markdown documentation
reqsmith ai generate-docs --format markdown --output api-docs.md

# OpenAPI specification
reqsmith ai generate-docs --format openapi --output api-spec.yaml

# Postman collection
reqsmith ai generate-docs --format postman --output collection.json

# HTML documentation
reqsmith ai generate-docs --format html --output docs/

Interactive AI Assistant​

Chat with AI​

Interactive AI assistance:

# Start AI chat session
reqsmith ai chat

# Ask specific questions
reqsmith ai ask "How do I handle pagination in this API?"

# Get help with specific error
reqsmith ai help --error "Connection timeout"

Chat Examples​

$ reqsmith ai chat

🤖 ReqSmith AI Assistant
Type 'exit' to quit, 'help' for commands

> How do I test rate limiting?

AI: To test rate limiting effectively:

1. Use burst requests:
   reqsmith repeat --count 100 --concurrent 10 api-endpoint

2. Monitor response headers:
   - X-RateLimit-Remaining
   - X-RateLimit-Reset
   - Retry-After

3. Test recovery:
   - Wait for rate limit reset
   - Implement exponential backoff

Would you like me to create a rate limit testing template?

> yes, create template for GitHub API rate limiting

AI: I'll create a comprehensive rate limiting test template...

AI Command Reference​

Core AI Commands​

# AI status and configuration
reqsmith ai status              # Check AI availability
reqsmith ai config              # Show AI configuration
reqsmith ai models              # List available AI models

# Request assistance
reqsmith ai suggest [URL]       # Get API endpoint suggestions
reqsmith ai analyze [ID]        # Analyze request/response
reqsmith ai diagnose [ID]       # Diagnose errors

# Template operations
reqsmith ai create-template     # Generate new template
reqsmith ai optimize-template   # Improve existing template
reqsmith ai analyze-template    # Analyze template quality

# Pattern recognition
reqsmith ai analyze-patterns    # Find usage patterns
reqsmith ai find-workflows      # Identify common workflows
reqsmith ai suggest-optimizations  # Performance suggestions

# Documentation
reqsmith ai generate-docs       # Create documentation
reqsmith ai generate-spec       # Create API specifications
reqsmith ai generate-examples   # Create usage examples

# Interactive assistance
reqsmith ai chat               # Start chat session
reqsmith ai ask [question]     # Ask specific question
reqsmith ai help [topic]       # Get help on topic

AI Options​

Common options for AI commands:

# Model selection
--model gemini-pro             # Specific AI model
--temperature 0.3              # Creativity level (0.0-1.0)

# Context options
--env production               # Include environment context
--template user-api            # Include template context
--history                      # Include request history

# Output options
--format json                  # Output format
--verbose                      # Detailed output
--quiet                        # Minimal output

# Analysis options
--focus security               # Focus on specific aspect
--include-logs                 # Include debug logs
--explain                      # Explain AI reasoning

Privacy and Security​

Data Handling​

Understanding how AI processes your data:

  • Local processing: Request metadata processed locally
  • API calls: Only anonymized data sent to AI service
  • No storage: AI provider doesn't store your data
  • Filtering: Sensitive data filtered before AI analysis

Security Configuration​

Configure AI security settings:

# Disable AI for sensitive environments
reqsmith config set ai.disabled_environments "production,staging"

# Filter sensitive data from AI analysis
reqsmith config set ai.filter_sensitive_data true

# Set sensitive headers to exclude
reqsmith config set ai.exclude_headers "Authorization,X-API-Key,Cookie"

# Anonymize URLs in AI requests
reqsmith config set ai.anonymize_urls true

Privacy Options​

Control what data is shared:

# Disable response body analysis
reqsmith config set ai.analyze_response_body false

# Disable URL analysis
reqsmith config set ai.analyze_urls false

# Use local AI model (when available)
reqsmith config set ai.prefer_local_model true

# Review AI requests before sending
reqsmith config set ai.review_before_send true

Advanced AI Features​

Custom Prompts​

Create custom AI prompts:

# Use custom analysis prompt
reqsmith ai analyze --prompt "Focus on REST API best practices"

# Create reusable prompt templates
reqsmith ai create-prompt \
  --name "security-review" \
  --template "Analyze this API response for security vulnerabilities"

# Use prompt templates
reqsmith ai analyze --prompt-template security-review

AI Workflows​

Combine AI features in workflows:

# Complete AI-assisted testing workflow
reqsmith ai workflow \
  --suggest-endpoints \
  --create-templates \
  --run-tests \
  --analyze-results \
  --generate-docs

AI Plugins​

Extend AI capabilities:

# List available AI plugins
reqsmith ai plugins list

# Install AI plugin
reqsmith ai plugins install security-scanner

# Enable plugin
reqsmith ai plugins enable security-scanner

Troubleshooting​

Common Issues​

  1. AI not responding: Check API key and network connection
  2. Poor suggestions: Provide more context in requests
  3. Rate limits: Monitor AI service usage limits
  4. Accuracy issues: Verify AI suggestions before using

Debug AI Features​

# Enable AI debugging
reqsmith config set ai.debug true

# Test AI connection
reqsmith ai test-connection

# Show AI request/response
reqsmith ai analyze --debug --explain

# Check AI service status
reqsmith ai service-status

Best Practices​

  1. Start simple: Begin with basic suggestions and analysis
  2. Provide context: Give AI relevant information for better results
  3. Verify suggestions: Always review AI recommendations
  4. Protect sensitive data: Configure appropriate privacy settings
  5. Monitor usage: Track AI service usage and costs
  6. Combine with manual testing: Use AI as an assistant, not replacement
  7. Keep learning: AI improves with more diverse usage patterns

Future AI Features​

Planned AI enhancements:

  • Local AI models: Offline AI capabilities
  • Custom model training: Train on your specific APIs
  • Advanced pattern recognition: More sophisticated usage analysis
  • Automated test generation: Complete test suite creation
  • Performance optimization: AI-driven performance tuning

Next Steps​