API Reference

Start Job

Start a previously created v1 job and pass model/background/composition settings.

Endpoint

POST /v1/jobs/{job_id}/start

Headers

HeaderRequiredDescription
X-Api-KeyYesAPI authentication
Content-Type: application/jsonYesJSON request body

Request Fields

FieldRequiredDescription
modelNooriginal | light | pro | human (uses job default when omitted)
text_promptNoUsed by pro model for subject targeting
backgroundNotransparent | color | image settings
compositionNoCanvas + foreground placement config
preview_durationNoPreview-only processing duration in seconds

Example Request

curl -X POST https://api.removebgvideo.com/v1/jobs/{JOB_ID}/start \
  -H "X-Api-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "pro",
    "text_prompt": "person, clothing accessories, glowing ring",
    "background": { "type": "transparent" },
    "composition": {
      "canvas": { "width": 1920, "height": 1080 },
      "foreground": {
        "anchor": "center",
        "scale": 1,
        "offset": { "x": 0, "y": 0 }
      }
    }
  }'

Response

{
  "success": true,
  "id": "job_123",
  "status": "processing",
  "message": "Job started"
}

API Contract Notes

  • All clients should handle non-2xx responses as structured error payloads.
  • Use explicit JSON schema validation for request payloads on your side.
  • Treat output_url as an asynchronous artifact and not an immediate response contract.
  • Persist job lifecycle state transitions for auditing and support.

Integration Verification

CheckHow to Verify
AuthenticationCall endpoint with valid and invalid key, confirm 200 vs 401
Rate limitsBurst test and confirm 429 handling with backoff
IdempotencyRetry same request and verify no duplicate side effects
ObservabilityConfirm request_id/job_id appears in logs and dashboards

Model/Feature Compatibility

Modeltext_promptBest Use Case
originalNoHigh quality general scenes
lightNoFast/simple scenes
proYesComplex object targeting
humanNoPortrait and people workflows

Server-side Validation Recommendations

  1. Reject unsupported model/output_format combinations early.
  2. Reject empty text_prompt when model=pro and prompt mode is required by your policy.
  3. Normalize composition defaults to avoid inconsistent render placement.