Setting OpenAI API Key Limits by Project

Set precise usage restrictions on OpenAI API keys by project to control budgets and model access:

1. Create a Project

• Sign in at platform.openai.com
• Navigate: Settings → Organization → Projects
• Click Create project, assign name (e.g., MyAppProject)

2. Generate Scoped API Key

• Open project dashboard, select API keys
• Click Create new secret key, assign identifier (e.g., ProjectXKey)
• Optional: Add IP allowlist for security
• Immediately copy generated key (one-time view only)

3. Configure Budget Limits

• Open Settings → Billing → Usage limits
• Set Hard limit (absolute cap) and Soft limit (warning threshold)

4. Restrict Model Usage

• Under project settings, go to Rate limits
• Allow specific models (e.g., gpt-4o-mini) with high limits
• Set zero limits for unwanted models, effectively disabling them

Additional Tips

• Regularly monitor usage via OpenAI’s dashboard
• Rotate keys periodically to maintain security
• Check API response headers for rate limit statuses (e.g., x-ratelimit-remaining-requests)

For advanced per-key controls, consider third-party or Azure OpenAI solutions.