API Reference¶
SIX exposes an OpenAI-compatible REST API. If you have code that talks to OpenAI, Anthropic, or any chat-completion provider, it works with SIX — and you get cryptographic receipts for free.
Base URL¶
Your endpoint is assigned during onboarding. All requests use HTTPS.
Authentication¶
Every request requires an Authorization header:
For high-assurance environments, cryptographic signature authentication is also available. See Authentication for details.
Endpoints¶
| Method | Path | Description |
|---|---|---|
POST |
/v1/chat/completions |
Run an inference and receive a signed receipt |
GET |
/v1/verify/{receipt_id} |
Verify a receipt independently |
Rate Limits¶
| Tier | Requests/min | Requests/day | Burst |
|---|---|---|---|
| Developer | 60 | 10,000 | 10 |
| Production | 600 | 100,000 | 50 |
| Enterprise | Custom | Custom | Custom |
Rate limit headers are returned with every response:
| Header | Description |
|---|---|
X-RateLimit-Limit |
Maximum requests per window |
X-RateLimit-Remaining |
Requests remaining in current window |
X-RateLimit-Reset |
Unix timestamp when the window resets |
Rate Limit Exceeded
When you exceed your rate limit, the API returns 429 Too Many Requests with a Retry-After header indicating how many seconds to wait.
Error Codes¶
| HTTP Code | Error Type | Description |
|---|---|---|
400 |
bad_request |
Malformed request body or missing required fields |
401 |
unauthorized |
Invalid or missing API key |
403 |
forbidden |
API key lacks permission for this operation |
404 |
not_found |
Resource (e.g., receipt) does not exist |
422 |
unprocessable_entity |
Valid JSON but invalid field values |
429 |
rate_limited |
Rate limit exceeded — see Retry-After header |
500 |
internal_error |
Server-side failure — request was not executed |
503 |
service_unavailable |
Temporary outage — safe to retry |
All error responses follow this format:
Response Format¶
All successful responses include:
- Standard OpenAI-compatible fields (
choices,model,usage) - A
receiptobject with cryptographic proof of execution
See Inference Endpoint for the full response schema.
Content Types¶
| Header | Value |
|---|---|
Content-Type |
application/json |
Accept |
application/json |
Next Steps¶
- Inference Endpoint -- Make verified inference requests
- Receipt Format -- Understand what each receipt field proves
- Verification Endpoint -- Verify receipts via API
- Integration Guides -- Language-specific examples