Skip to main content

​
Base URL

https://api.trainlyai.com
All endpoints are scoped to a project: 
https://api.trainlyai.com/v1/{project_id}/...

​
Authentication

Include your API key in the Authorization header. Keys are prefixed with tk_. 
Authorization: Bearer tk_your_api_key

​
Rate Limiting

All endpoints are rate-limited to 100 requests per second per API key. Exceeding this limit returns a 429 response with a Retry-After header.

​
Response Format

All responses return JSON with Content-Type: application/json.

​
Error Codes

CodeDescription
400Bad Request — invalid or missing parameters
401Unauthorized — missing or invalid API key
403Forbidden — API key lacks permission for this resource
404Not Found — resource does not exist
429Too Many Requests — rate limit exceeded
500Internal Server Error — unexpected failure
Error responses include a JSON body: 
{
  "error": {
    "code": 400,
    "message": "input is required"
  }
}

​
API Categories

CategoryDescription
TracesLog, list, and inspect AI call traces
ScoringAttach scores to traces manually or via LLM judges
AnalyticsQuery metrics, costs, performance, and export logs
TestingCreate test suites, run evaluations, view results
VersionsPublish, compare, and rollback prompt/model versions

​
Quick Start

Log a trace with a single cURL call: 
curl -X POST https://api.trainlyai.com/v1/proj_abc123/traces \
  -H "Authorization: Bearer tk_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "Summarize this document.",
    "output": "The document describes quarterly revenue growth of 12%.",
    "model": "gpt-4o",
    "latency_ms": 842.5,
    "token_usage": {
      "prompt_tokens": 320,
      "completion_tokens": 45,
      "total_tokens": 365
    },
    "tags": ["summarization", "production"],
    "status": "success"
  }'
Response: 
{
  "trace_id": "trc_9f8e7d6c5b4a",
  "status": "created"
}