Skip to content

API Keys Setup

Jump to bottom
thejaustin edited this page Dec 25, 2025 · 1 revision

API Keys Setup Guide

Complete guide to obtaining and configuring LLM provider API keys for AutoCat.

Overview

AutoCat requires an API key from at least one LLM provider to enable AI-powered categorization. This guide covers:

  1. How to obtain API keys from each provider
  2. How to configure keys in AutoCat
  3. Cost and rate limit information
  4. Security best practices

🔑 Provider Options

Google AI (Recommended for Free Tier)

Best for: Free tier users, fast categorization, generous limits

Pros:

  • Free tier with generous limits
  • Fast response times (Gemini 2.0 Flash)
  • Large context window (1M tokens)
  • No credit card required for free tier

Cons:

  • Rate limited (15 requests per minute on free tier)

Cost: Free tier available

Claude (Anthropic)

Best for: High accuracy, detailed reasoning

Pros:

  • Excellent categorization accuracy
  • Detailed reasoning output
  • Fast models (Claude 3.5 Haiku)

Cons:

  • Requires credit card
  • Paid service (no free tier)
  • More expensive than Google AI

Cost: Pay-as-you-go (starts at $0.25 per million input tokens)

OpenAI

Best for: Balanced performance

Pros:

  • Well-established provider
  • Good categorization accuracy
  • Fast models (GPT-4o Mini)

Cons:

  • Requires credit card
  • Paid service (no free tier)
  • Rate limits on free trial

Cost: Pay-as-you-go (starts at $0.15 per million input tokens)

Perplexity

Best for: Online search integration

Pros:

  • Online search capabilities
  • Fast models (Llama 3.1 Sonar)
  • Good for research-based categorization

Cons:

  • Requires credit card
  • Paid service (limited free tier)
  • Less common provider

Cost: Free tier with $5 credit, then pay-as-you-go

🔧 Setup Instructions

Option 1: Google AI (Gemini)

Step 1: Get API Key

  1. Go to Google AI Studio
  2. Click "Get API Key" or "Get Started"
  3. Sign in with Google account
  4. Click "Create API Key"
  5. Select or create a Google Cloud project
  6. Copy the API key (starts with AIza...)

Step 2: Configure in AutoCat

Method A: Settings UI

  1. Open AutoCat
  2. Settings → Categorization → LLM Settings
  3. Scroll to "Google AI" section
  4. Tap "API Key" field
  5. Paste your API key
  6. Tap "Test Connection" to verify

Method B: Environment Variable (More secure)

# In Termux or shell
export GOOGLE_AI_API_KEY="AIzaSy..."

# AutoCat will automatically detect this

Step 3: Select Model

  1. Settings → Categorization → LLM Settings
  2. Under "Google AI", tap "Model"
  3. Select "Gemini 2.0 Flash (Recommended)"
  4. Save settings

Rate Limits

  • Free tier: 15 requests per minute
  • Quota: 1,500 requests per day
  • Batch processing: Recommended to reduce requests

Option 2: Claude (Anthropic)

Step 1: Get API Key

  1. Go to Anthropic Console
  2. Sign up or log in
  3. Click "API Keys" in sidebar
  4. Click "Create Key"
  5. Give it a name (e.g., "AutoCat")
  6. Copy the key (starts with sk-ant-...)

Note: Requires credit card on file

Step 2: Configure in AutoCat

Method A: Settings UI

  1. Open AutoCat
  2. Settings → Categorization → LLM Settings
  3. Tap "LLM Provider" → Select "Claude"
  4. Tap "API Key" field under "Claude"
  5. Paste your API key
  6. Tap "Test Connection" to verify

Method B: Environment Variable

export ANTHROPIC_API_KEY="sk-ant-..."

Step 3: Select Model

Recommended models:

  • Claude 3.5 Haiku: Fast, affordable
  • Claude 3.5 Sonnet: More accurate, slower

Rate Limits

  • Tier 1 (new accounts): 50 requests per minute
  • Tier 2 (after $5 spend): 1,000 requests per minute

Option 3: OpenAI

Step 1: Get API Key

  1. Go to OpenAI Platform
  2. Sign up or log in
  3. Click on your profile → "API Keys"
  4. Click "Create new secret key"
  5. Name it (e.g., "AutoCat")
  6. Copy the key (starts with sk-...)

Note: Requires credit card on file

Step 2: Configure in AutoCat

Method A: Settings UI

  1. Open AutoCat
  2. Settings → Categorization → LLM Settings
  3. Tap "LLM Provider" → Select "OpenAI"
  4. Tap "API Key" field under "OpenAI"
  5. Paste your API key
  6. Tap "Test Connection" to verify

Method B: Environment Variable

export OPENAI_API_KEY="sk-..."

Step 3: Select Model

Recommended model:

  • GPT-4o Mini: Fast, affordable, good accuracy

Rate Limits

  • Free trial: 3 requests per minute
  • Tier 1 (after $5 spend): 500 requests per minute

Option 4: Perplexity

Step 1: Get API Key

  1. Go to Perplexity AI
  2. Sign up or log in
  3. Navigate to API settings
  4. Create new API key
  5. Copy the key

Step 2: Configure in AutoCat

Method A: Settings UI

  1. Open AutoCat
  2. Settings → Categorization → LLM Settings
  3. Tap "LLM Provider" → Select "Perplexity"
  4. Tap "API Key" field under "Perplexity"
  5. Paste your API key
  6. Tap "Test Connection" to verify

Method B: Environment Variable

export PERPLEXITY_API_KEY="pplx-..."

Step 3: Select Model

Recommended model:

  • Llama 3.1 Sonar Small: Fast, affordable

💰 Cost Comparison

For 100 Apps Categorization (with batch processing)

Provider Model Estimated Cost Free Tier
Google AI Gemini 2.0 Flash $0.00 ✅ Free
Claude Claude 3.5 Haiku ~$0.01-0.02 ❌ Paid
OpenAI GPT-4o Mini ~$0.01 ❌ Paid
Perplexity Llama 3.1 Sonar ~$0.01 ⚠️ $5 credit

Note: Costs are estimates based on typical app metadata. Actual costs may vary.

Monthly Cost Estimates

Assuming daily re-categorization of 100 apps:

Provider Daily Monthly (30 days)
Google AI Free Free
Claude $0.01 $0.30
OpenAI $0.01 $0.30
Perplexity Free (with credit) $0.30 (after credit)

🔒 Security Best Practices

Use Environment Variables

Why?

  • API keys not stored in app settings
  • More secure (not in backups)
  • Easy to rotate keys

How?

Termux:

# Add to ~/.bashrc or ~/.zshrc
export GOOGLE_AI_API_KEY="your-key-here"
export ANTHROPIC_API_KEY="your-key-here"
export OPENAI_API_KEY="your-key-here"
export PERPLEXITY_API_KEY="your-key-here"

# Reload shell
source ~/.bashrc

Android (via Termux):

# Create a script to set env vars
cat > ~/set-api-keys.sh <<EOF
export GOOGLE_AI_API_KEY="your-key-here"
export ANTHROPIC_API_KEY="your-key-here"
export OPENAI_API_KEY="your-key-here"
export PERPLEXITY_API_KEY="your-key-here"
EOF

# Source before running AutoCat
source ~/set-api-keys.sh

Key Rotation

Best practice: Rotate API keys every 90 days

  1. Generate new key in provider console
  2. Update AutoCat settings or env variable
  3. Test connection
  4. Delete old key in provider console

Monitor Usage

Prevent unexpected charges:

  1. Set up billing alerts in provider console
  2. Monitor API usage regularly
  3. Set spending limits if available
  4. Review AutoCat accuracy metrics (Settings → LLM Settings)

🧪 Testing Your Configuration

Test Connection Button

After adding API key:

  1. Settings → Categorization → LLM Settings
  2. Find your provider section
  3. Tap "Test Connection"
  4. Wait for result:
    • ✅ Success: Shows latency and model version
    • ❌ Failure: Shows specific error message

Common Test Errors

"Invalid API key"

  • Double-check key is copied correctly
  • Verify key is active in provider console
  • Check for extra spaces

"Rate limit exceeded"

  • Wait a few seconds and try again
  • Check your provider's rate limits
  • Consider upgrading tier

"Network error"

  • Check internet connection
  • Verify provider API is not down
  • Check firewall/VPN settings

📊 Choosing the Right Provider

Decision Matrix

Use Case Recommended Provider Why
Just trying AutoCat Google AI Free, no credit card needed
Daily use (free) Google AI Generous free tier
Best accuracy Claude 3.5 Haiku Higher quality categorization
Balanced (cost/quality) OpenAI GPT-4o Mini Good quality, reasonable cost
Auto-select best Enable "Auto-Select" Let AutoCat choose based on accuracy

Auto-Select Best Model

What it does: AutoCat tracks accuracy for each provider and automatically selects the best performing one.

How to enable:

  1. Settings → Categorization → LLM Settings
  2. Enable "Auto-Select Best Model"
  3. Ensure you have API keys for multiple providers
  4. AutoCat will analyze 30 days of accuracy data

Requirements:

  • At least 2 providers configured
  • Minimum 10 categorizations per provider
  • At least 70% accuracy threshold

🔄 Switching Providers

Mid-Use Provider Change

You can switch providers at any time:

  1. Settings → Categorization → LLM Settings
  2. Tap "LLM Provider"
  3. Select new provider
  4. Ensure new provider has valid API key
  5. Test connection

Note: Existing categorizations remain unchanged. Only new categorizations use new provider.

Re-categorize with New Provider

To re-categorize all apps with new provider:

  1. Switch provider (above steps)
  2. Settings → Categorization
  3. Tap "Re-categorize All Apps"
  4. Wait for completion

❓ FAQ

Q: Can I use multiple providers? A: Yes! Add keys for multiple providers and enable "Auto-Select Best Model" to let AutoCat choose.

Q: Do I need to pay for AutoCat? A: No, AutoCat is free. You only pay for LLM provider API usage (Google AI has free tier).

Q: How many apps can I categorize for free? A: With Google AI free tier: ~1,500 categorizations per day (more with batch processing).

Q: What if my key stops working? A: Check provider console for:

  • Key status (active/revoked)
  • Billing issues
  • Rate limit exceeded
  • API quota exhausted

Q: Can I share my API key? A: No, API keys are personal. Sharing violates provider terms of service.

Q: How do I delete my API key from AutoCat? A: Settings → Categorization → LLM Settings → Tap API key field → Delete text

🆘 Troubleshooting

"Connection failed" Error

  1. Check internet connection
  2. Verify API key is correct (copy/paste again)
  3. Test in provider console (make test API call)
  4. Check provider status (is API down?)

"Quota exceeded" Error

  1. Check provider dashboard for usage limits
  2. Wait for quota reset (usually daily/hourly)
  3. Upgrade tier if available
  4. Switch to different provider temporarily

"Invalid model" Error

  1. Update AutoCat to latest version
  2. Check model availability in provider docs
  3. Select different model in settings
  4. Report issue if model should be available

📚 Additional Resources

Provider Documentation

AutoCat Documentation

Last Updated: 2025-12-25

AutoCat | Fork of Lawnchair 15 | Download Latest Build | Last Updated: 2025-12-25

🏠 Navigation

Getting Started

Documentation

Development

Planning

🔗 External Links

Clone this wiki locally