Frequently Asked Questions
Getting Started
Q: How do I get an API key?
A: Visit
idlecloud.ai and click "Get Started". Complete the authentication flow and onboarding process. You'll then be able to generate an API key from your dashboard.
Q: How do I install the Python SDK?
A: Run pip install idlecloud in your terminal.
Q: Is IdleCloud compatible with OpenAI?
A: Yes! Our API is 100% compatible with OpenAI's interface. You can use either our SDK or OpenAI's SDK with our base URL (https://api.idlecloud.ai/v1).
Billing & Pricing
Q: What are the rate limits?
A: Rate limits vary by usage tier. Check the X-RateLimit-* headers in API responses for your current limits. We're still finalizing our rate limit policies during beta.
Q: How am I billed?
A: You're billed based on tokens used. Pricing is per million tokens. See
our pricing page for current rates.
Q: Do you offer free credits?
A: New users receive $1 in free credits upon signup to test the service.
Technical
Q: What models are available?
A: Currently we support gpt-oss-20b. More models coming soon, including gpt-oss-20b-safeguard for moderation tasks.
Q: Do you support streaming?
A: Yes! Set stream=True in your request to receive streamed responses.
Q: What are reasoning tokens?
A: The gpt-oss-20b model uses internal reasoning before generating answers. These reasoning tokens are included in completion_tokens but not shown in the response. You are billed for all tokens including reasoning tokens.
Q: How do I handle rate limits?
A: Check the
X-RateLimit-* headers in responses. Implement exponential backoff for 429 errors. See our
documentation for code examples.
Account Management
Q: How do I reset my API key?
A: Log in to your dashboard and click "Regenerate API Key". Old keys are immediately invalidated.
Q: Can I have multiple API keys?
A: Yes, you can create as many as you want from your dashboard.
Security & Privacy
Q: Is my data secure?
A: Yes. All API requests use HTTPS encryption. We don't store your prompts or responses.
Q: If my requests are being processed on private computers, how do I know it's safe?
A: We use a novel and proprietary encryption scheme that enables your requests to be run openly on our miners' GPUs while being totally unreadable.
Q: Do you train models on my data?
A: No. Your API requests are never used for training.
Q: Where are your servers located?
A: Our network is decentralized across global miner nodes. Data is processed on nodes closest to you for optimal latency.