Async jobs

Some generate endpoints can enqueue work and return immediately. Poll the job until it finishes.

Get job status

GET /v1/ai/generate/job/:jobId

curl https://api.krastie.ai/v1/ai/generate/job/123 \
  -H "api-key: YOUR_KEY"

Success (200)

{
  "jobId": "123",
  "name": "ai_tool_content",
  "state": "completed",
  "progress": 100,
  "result": {},
  "failedReason": null,
  "attemptsMade": 1,
  "processedOn": 1710000000000,
  "finishedOn": 1710000005000,
  "data": {},
  "queue": "aiQueue"
}

Field	Description
state	Queue state such as waiting, active, completed, failed
progress	Progress value when available
result	Job return value when completed
failedReason	Failure message when the job failed
queue	Which internal queue held the job (aiQueue, seoQueue, or contentQueue)

Not found (404)

{ "message": "Job not found", "jobId": "123" }

Queuing text generation

curl https://api.krastie.ai/v1/ai/generate/text \
  -H "api-key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "text": "Draft a landing page outline.",
    "useWorker": true
  }'

Response (202):

{
  "message": "Content generation queued",
  "jobId": "123"
}

Poll until state is completed or failed. Respect rate limits if you poll frequently.