Each API request costs credits based on the capabilities used. Credits pay for orchestration, tool execution, sandboxing, and file storage — your own LLM key covers the AI inference (see BYOK).
Capability
Credits
Description
Chat
1
Conversational AI
Data
1
Structured data processing
Files
2
File upload/download via signed URLs
Research
3
Web search + synthesis
Code
3
Sandboxed code execution
Documents
5
PDF/report generation
A request using multiple capabilities costs the sum of their weights. For example, a request with capabilities: ["research", "documents"] costs 8 credits.
Credits are pre-deducted before your request is processed. If the compute service fails, credits are automatically refunded to your balance.When your balance reaches zero, requests return a 402 status:
{ "error": { "code": "insufficient_credits", "message": "No credits remaining. Purchase credits at https://www.stewrd.dev/pricing" }}
All requests are rate-limited to 60 requests per minute.If you exceed this, you’ll receive a rate_limited error with HTTP status 429. Wait and retry with exponential backoff.