> ## Documentation Index
> Fetch the complete documentation index at: https://docs.clavata.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# October 6, 2026

> Job feedback can report missing frames and malformed transcripts.

**October 6, 2026**

What's new

## Improvements

### Report extraction failures with job feedback

[Submit job feedback](https://docs.clavata.ai/api-reference/job-feedback/submit-job-feedback.md)
(`POST /v1/jobs/{jobUuid}/feedback`) items now take an optional `kind`, so you
can report a problem with the content a label was evaluated on rather than with
the label's outcome:

* `JOB_FEEDBACK_KIND_MISCLASSIFICATION`: the label's outcome was wrong.
  `expectedOutcome` is required. Omitting `kind` means the same thing, so
  existing integrations keep working unchanged.
* `JOB_FEEDBACK_KIND_MISSING_FRAME`: a video frame that should have been
  extracted was missing.
* `JOB_FEEDBACK_KIND_MALFORMED_TRANSCRIPT`: the video's transcript was
  malformed.

`expectedOutcome` is now optional, and must be omitted for the two extraction
kinds, since there is no label outcome to disagree with. Extraction feedback is
also accepted on a label that did not reach a definite result.
[Get job feedback](https://docs.clavata.ai/api-reference/job-feedback/get-job-feedback.md)
(`GET /v1/jobs/{jobUuid}/feedback`) returns each entry's `kind`, with
`expectedOutcome` and `actualOutcome` set to `EVALUATION_RESULT_UNSPECIFIED`
where they don't apply.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.