GET /v1/jobs/{job_id}

Poll an async scoring job.

GET /v1/jobs/{job_id}

Poll an async scoring job.

cURL

curl --request GET \
  --url https://api.keplerinsights.us/v1/jobs/{job_id} \
  --header 'X-API-Key: <api-key>'

200

{
  "mode": "<string>",
  "job_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
  "status": "pending",
  "domain": "<string>",
  "created_at": "2023-11-07T05:31:56Z",
  "completed_at": "2023-11-07T05:31:56Z",
  "result_ref": {
    "domain": "<string>",
    "scored_at": "2023-11-07T05:31:56Z"
  },
  "score_url": "/v1/score/stripe.com",
  "failure_reason": "sfn_failed"
}

Returns the current state of a cold-scoring job. Jobs are created automatically by POST /v1/score when no fresh cached record exists.

Status transitions

status	Meaning
pending	The underlying cold pipeline is still running. Poll again in ~5s.
complete	The pipeline succeeded. result_ref and score_url are populated.
failed	The pipeline ended in a non-success terminal state. failure_reason explains.

Status is settled lazily — the first poll that observes a terminal underlying execution flips the row. A job that’s never polled stays pending in storage until its 30-day TTL.

Reading a complete job

{
  "job_id":       "9f0c2d83-...",
  "status":       "complete",
  "domain":       "stripe.com",
  "created_at":   "2026-05-12T18:30:00Z",
  "completed_at": "2026-05-12T18:30:43Z",
  "result_ref":   { "domain": "stripe.com", "scored_at": "2026-05-12T18:30:43Z" },
  "score_url":    "/v1/score/stripe.com"
}

Once you have the score_url, do GET https://api.keplerinsights.us/v1/score/stripe.com to read the actual score. The two-step is intentional: the job row stays small and cheap; the full score response stays the same shape regardless of how it was triggered.

Failure reasons

failure_reason	Meaning
sfn_failed	The state-machine ended in FAILED (engine raised an exception).
sfn_timed_out	Step Functions hit its task timeout. Usually transient — retry.
sfn_aborted	An operator manually aborted the run. Should never happen in production.
scoring_completed_no_record	The state-machine succeeded but no score-history row appeared. Possible cause: insufficient data stub. Check /v1/company/{domain}/confidence.

Cross-user access

Job IDs are server-side bound to the API key that created them. Polling someone else’s job ID returns:

HTTP/1.1 403 Forbidden
{ "error": "forbidden" }

If you’re proxying multiple downstream users through one Kepler key, you’ll need to map their session IDs to your job_ids yourself.

Expiration

Jobs are kept 30 days then deleted via DynamoDB TTL. Polling an expired job returns 404. Don’t rely on job rows for long-term history; use GET /v1/score/{domain}/history for that.

Authorizations

X-API-Key · string · header · required

Live keys are prefixed ki_live_, test keys ki_test_. Issue + revoke keys at https://api.keplerinsights.us. Never embed a key in client-side code — every endpoint is backend-to-API only.

Path Parameters

job_id · string<uuid> · required

Response

200

application/json

Job state.

mode · string

job_id · string<uuid>

status · enum<string>

Available options: pending, complete, failed

domain · string

created_at · string<date-time>

completed_at · string<date-time> | null

result_ref · object

Present on complete. Use score_url to fetch the actual record.

Show child attributes

score_url · string | null

Example:

"/v1/score/stripe.com"

failure_reason · enum<string> | null

Available options: sfn_failed, sfn_timed_out, sfn_aborted, scoring_completed_no_record