Skip to content

API Keys

Configure AI provider API keys to enable analysis capabilities. Multiple keys provide faster performance through automatic load balancing and higher rate limits.

  • Faster Analysis: Multiple keys mean higher combined rate limits
  • Better Reliability: Automatic failover if one provider has issues
  • Load Distribution: Spreads requests across providers to avoid throttling
  • Optimal Selection: Uses the best model for each specific task automatically

Add as many as you have:

  • OpenAI: GPT models for general analysis
  • Anthropic: Claude models for complex reasoning
  • Google: Gemini models for multimodal analysis

The more providers you configure, the faster and more reliable your analysis becomes.

  1. Open Settings by clicking the gear icon in the top navigation
  2. Go to API Keys section
  3. Add your keys for any providers you have accounts with
  4. Save - Probably handles everything else automatically

No Configuration Needed

You don’t need to choose models, adjust settings, or make technical decisions. Probably automatically uses the best available model for each task and balances load across your keys.

OpenAI

Starts with sk-

sk-1234567890abcdef…

Anthropic

Starts with sk-ant-

sk-ant-1234567890…

Google

Starts with AIza

AIzaSyA1234567890…

OpenAI

  1. Visit platform.openai.com
  2. Create account or log in
  3. Go to API Keys section
  4. Create new secret key
  5. Copy and paste into Probably

Anthropic Claude

  1. Visit console.anthropic.com
  2. Create account or log in
  3. Go to API Keys
  4. Generate new key
  5. Copy and paste into Probably

Google Gemini

  1. Visit aistudio.google.com
  2. Create account or log in
  3. Get API key
  4. Copy and paste into Probably

Probably intelligently routes different tasks to the most appropriate models:

  • Text Analysis: Routes to models best at language understanding
  • Data Reasoning: Uses models optimized for analytical thinking
  • Multimodal Tasks: Automatically uses models that can handle images/documents
  • Load Balancing: Distributes requests to avoid rate limits

You never need to:

  • Choose which model to use
  • Decide between providers
  • Adjust technical settings
  • Monitor performance metrics

Just add your keys and let Probably handle the optimization.

✅ Multiple Providers

  • Faster analysis through parallel processing
  • Higher combined rate limits
  • Automatic failover for reliability
  • Best model selection for each task

⚠️ Single Provider

  • Limited by one provider’s rate limits
  • Potential delays during peak usage
  • No fallback if provider has issues
  • Suboptimal model selection
  • Agent Responses: Faster chat responses with multiple keys
  • Large Dataset Analysis: Better performance on complex data
  • Concurrent Users: Multiple team members can work simultaneously
  • Peak Hours: Avoid slowdowns during busy periods
  1. Add Multiple Providers: Get keys from OpenAI, Anthropic, and Google for best performance
  2. Keep Keys Current: Replace expired or revoked keys promptly
  3. Monitor Usage: Check your provider dashboards occasionally for usage patterns
  4. Secure Storage: Never share your API keys or commit them to code repositories
  1. Start Simple: Begin with one API key to get familiar with Probably
  2. Add More Keys: As you use Probably more, add additional providers
  3. Team Usage: For teams, ensure multiple keys are configured for concurrent access
  4. Peak Times: Add extra keys if you notice slowdowns during busy periods
  • Local Encryption: All API keys are encrypted and stored locally on your machine
  • No External Transmission: Keys are never sent to external servers except for authorized API calls
  • Secure Storage: Industry-standard encryption protects your credentials
  • Privacy Compliance: Full compliance with data privacy regulations
  • User-Level Storage: Keys are stored per user account
  • No Sharing: Keys are never shared between users or transmitted externally
  • Revocation: You can remove keys anytime from the settings interface
  • Audit Trail: Clear visibility into which keys are active

Performance Optimization

Learn how to optimize performance and understand the impact of your configuration.

Troubleshooting

Resolve common API key issues and configuration problems.