Rubric video evaluation

Evaluations: Score videos against configurable rubrics with reviewable evidence.

Upload a video, choose a rubric, and get a per-item evaluation table with credited / not credited / needs review outcomes, confidence, timestamped evidence, screenshots, and derived scores. Reviewers can override each AI call with an audit note, making this the general-purpose rubric engine for education, clinical training, sales QA, compliance, and other evaluation workflows.

Built for Education, clinical training, compliance, coaching, and QA teams.

Sample output

Rubric score table

Item: asks open-ended question — credited
Evidence: 02:11 timestamp and screenshot
Score: 8/10 composite

What to upload

A video to evaluate plus a rubric template.

  • Early-childhood classroom observation
  • Clinical OSCE or practical-skills recording
  • Sales-call quality review
  • Retail or safety compliance walkthrough

What you get

A review table with item-level AI outcomes, evidence, screenshots, overrides, and computed scores.

  • Per-rubric-item status and confidence
  • Timestamped evidence text
  • Screenshot from the evidence moment
  • Derived scores computed from rubric formulas
  • Human override audit trail

Exports

  • CSV (one row per rubric item)
  • JSON (full evaluation payload)

Tips for better results

Up to 1 hour per evaluation by default
Use stable rubric item IDs before running production evaluations
Put age, role, cohort, or site fields in subject metadata so gates can apply
Treat overrides as calibration data, not just edits