Skip to main content

​
Overview

Parsaa supports multiple AI providers, allowing you to choose the best model for your specific needs. Configure your API keys to access different AI services and models.
Privacy First: All API keys are stored locally on your machine and never transmitted to Parsaa servers.

​
Supported Providers

​
OpenAI

  1. Visit OpenAI Platform
  2. Sign in to your OpenAI account
  3. Click “Create new secret key”
  4. Copy the key and add it to Parsaa
  • GPT-4: 0.03/1Kinputtokens,0.03/1K input tokens, 0.06/1K output tokens
  • GPT-3.5 Turbo: 0.001/1Kinputtokens,0.001/1K input tokens, 0.002/1K output tokens
  • GPT-4 Turbo: 0.01/1Kinputtokens,0.01/1K input tokens, 0.03/1K output tokens

​
Anthropic (Claude)

  1. Visit Anthropic Console
  2. Sign in to your Anthropic account
  3. Navigate to “API Keys” section
  4. Create a new key and add it to Parsaa
  • Claude-3 Opus: 15/1Minputtokens,15/1M input tokens, 75/1M output tokens
  • Claude-3 Sonnet: 3/1Minputtokens,3/1M input tokens, 15/1M output tokens
  • Claude-3 Haiku: 0.25/1Minputtokens,0.25/1M input tokens, 1.25/1M output tokens

​
Google (Gemini)

  1. Visit Google AI Studio
  2. Sign in with your Google account
  3. Create a new API key
  4. Add the key to Parsaa settings
  • Gemini Pro: Free tier available, then $0.50/1M input tokens
  • Gemini Pro Vision: $0.50/1M input tokens
  • Gemini Ultra: $1.25/1M input tokens

​
Fireworks AI

  1. Visit Fireworks AI
  2. Create an account and sign in
  3. Navigate to API Keys section
  4. Generate a new key for Parsaa
  • Llama 2 70B: $0.40/1M input tokens
  • Code Llama 34B: $0.20/1M input tokens
  • Mixtral 8x7B: $0.15/1M input tokens

​
Voyage AI

  1. Visit Voyage AI
  2. Sign up for an account
  3. Access your API dashboard
  4. Create a new API key
  • Code Embeddings: For semantic code search
  • Similarity Search: Find similar code patterns
  • Code Clustering: Group related code together
  • Voyage-2: $0.10/1M input tokens
  • Voyage-3: $0.20/1M input tokens
  • Custom Models: Contact for pricing

​
OpenRouter

  1. Visit OpenRouter
  2. Create an account
  3. Generate an API key
  4. Add it to Parsaa for access to multiple models
  • Access to 100+ models from various providers
  • Unified API for different model providers
  • Cost comparison and optimization
  • Model Switching: Easy switching between providers
  • Cost Optimization: Choose the most cost-effective model
  • Fallback Support: Automatic fallback to backup models

​
Adding API Keys to Parsaa

​
Method 1: Settings Panel

1

Open Settings

In Xcode, go to Parsaa > Settings or use ⌘,
2

Navigate to API Keys

Click on the “API Keys” tab in the settings panel
3

Add Provider

Click “Add Provider” and select your AI provider
4

Enter API Key

Paste your API key and click “Save”
5

Test Connection

Click “Test” to verify the API key works correctly

​
Method 2: Command Line

# Add OpenAI key
parsaa config set openai.api_key "your-key-here"

# Add Claude key
parsaa config set claude.api_key "your-key-here"

# List all configured keys
parsaa config list

# Set environment variables
export OPENAI_API_KEY="your-key-here"
export CLAUDE_API_KEY="your-key-here"
export GOOGLE_API_KEY="your-key-here"

# Parsaa will automatically detect these

​
Model Selection Strategy

​
For Code Analysis

Complex Analysis

Recommended: GPT-4 or Claude-3 Opus
  • Best for understanding complex code
  • Excellent reasoning capabilities
  • Higher cost but better results

Quick Analysis

Recommended: GPT-3.5 Turbo or Claude-3 Haiku
  • Fast response times
  • Cost-effective for simple tasks
  • Good for basic code review

​
For Code Generation

Swift/SwiftUI

Recommended: GPT-4 or Code Llama
  • Specialized in Swift development
  • Understands iOS patterns
  • Best for UI code generation

General Purpose

Recommended: Claude-3 Sonnet or Gemini Pro
  • Good balance of quality and cost
  • Versatile for different languages
  • Reliable performance

​
For Code Refactoring

Complex Refactoring

Recommended: GPT-4 or Claude-3 Opus
  • Understands code relationships
  • Preserves functionality
  • Suggests architectural improvements

Simple Refactoring

Recommended: GPT-3.5 Turbo or Claude-3 Haiku
  • Quick style improvements
  • Basic optimization
  • Cost-effective for simple changes

​
Cost Optimization

​
Smart Model Selection

Parsaa can automatically choose the best model based on:
  • Task Complexity: Simple tasks use cheaper models
  • Code Size: Large files use more capable models
  • Context Requirements: Complex context needs advanced models
  • Cost Budget: Set limits to control spending
You can always override automatic selection:
  • Force Specific Model: Use a particular model for all tasks
  • Cost Limits: Set maximum cost per request
  • Quality Thresholds: Minimum quality requirements
  • Fallback Models: Backup options if primary fails

​
Usage Monitoring

  • Real-time Usage: See costs as you use Parsaa
  • Daily Limits: Set spending limits per day
  • Monthly Budgets: Track monthly AI costs
  • Usage Reports: Detailed breakdown of expenses
  • Batch Requests: Group similar requests together
  • Context Optimization: Reduce unnecessary context
  • Model Switching: Use cheaper models for simple tasks
  • Caching: Cache similar requests to avoid duplicates

​
Security Best Practices

​
API Key Security

Never share your API keys. Keep them secure and never commit them to version control.
  • Local Storage: Keys stored locally on your machine
  • Encryption: Keys encrypted at rest
  • No Transmission: Keys never sent to Parsaa servers
  • Access Control: Only you can access your keys
  • Regular Rotation: Change keys periodically
  • Limited Scope: Use keys with minimal required permissions
  • Monitoring: Track key usage and access
  • Revocation: Ability to revoke compromised keys

​
Privacy Considerations

  • Code Analysis: Your code is sent to AI providers for analysis
  • No Storage: AI providers don’t store your code permanently
  • Usage Policies: Review each provider’s data usage policies
  • Sensitive Code: Consider using local models for sensitive projects
  • Ollama: Run models locally for complete privacy
  • LM Studio: Local model management
  • Custom Models: Use your own fine-tuned models
  • Hybrid Mode: Combine local and cloud processing

​
Troubleshooting

​
Common Issues

Symptoms: “Invalid API key” errorsSolutions:
  1. Verify the key is correct and complete
  2. Check if the key has expired
  3. Ensure the key has proper permissions
  4. Try regenerating the key
Symptoms: “Rate limit exceeded” errorsSolutions:
  1. Wait for the rate limit to reset
  2. Upgrade to a higher tier plan
  3. Use a different model provider
  4. Implement request queuing
Symptoms: “Model not available” errorsSolutions:
  1. Check if the model is supported
  2. Verify your account has access
  3. Try a different model
  4. Contact the provider for support

​
Getting Help

Contact Support

Having trouble with API keys? Our support team is here to help.
Beta Support: During the beta period, we provide priority support for all API key and configuration issues.