Test a Video or Image

Include a camera name only if your policy is influenced by camera/location name.
Policy text override in place. Selected policies will not be used.
Use this to OVERRIDE your saved policy. Can be written in either plain text or Markdown. Links and media will be ignored. Use the Preview tab to view how your Markdown formatting will be interpreted.

Define zones of interest to extract from uploaded media.

  • Behavior: Include crops and submits the zone region for policy evaluation; Exclude suppresses detections inside it and excludes the region from Tracing crops; Blackout redacts the region in derivative media, leaving originals unchanged; Inscribe overlays a label or border onto derivative media.
  • Camera: Restrict the zone to a specific camera by name, or apply it to all cameras.
  • X, Y, W, H: Position and size as percentages of source media dimensions (0–100). X/Y is the top-left corner; W/H is width and height.
  • Video / Images: (Include only) Select what to extract from the zone region. Video requires a video source; a 13.5MiB inline limit applies, preprocessing handles larger files remotely.
  • FPS: (Include only) Frames per second to extract from source video. Leave blank for default.
  • Apply to Include Zones: (Blackout / Inscribe) Whether the transform is also applied within Include zone crops.
  • Include Full Frame Derivative: Include the full frame with all blackout and inscribe transforms applied for policy evaluation.
  • Apply to Derivative Media: (Blackout / Inscribe) Whether the transform is applied to the derivative media artifact.
  • Apply to Tracing: (Blackout / Inscribe) Whether the transform is applied to Tracing crop outputs.
Includes the original media for policy evaluation alongside zone and tracing crops.
Includes transformed media with blackout and inscribe zones applied for policy evaluation.
Includes the full frame with all transforms applied for policy evaluation.
Zone
Camera
Applied to all cameras
Coordinates
X %
Y %
W %
H %
[[ {'include': 'Extraction', 'exclude': 'Effect', 'blackout': 'Redaction', 'inscribe': 'Inscription'}[zone.behavior || 'include'] || 'Extraction' ]]

Upload disabled for [[ zone.behavior ]] zones.

No zones defined.
Preview
[[ preview.name ]] [[ preview.width ]] x [[ preview.height ]]px

[[ preview.errorMsg ]]

Browser Support Tips:
  • Videos: Use MP4 (H.264/AAC) or WebM. H.265/HEVC is often not supported.
  • Images: JPG, PNG, WebP, and GIF are universally supported.
Detects and tracks objects in video clips, generating per-object crops from the original footage.
Uploads the original media for policy evaluation in addition to tracing crops.
Includes transformed media with blackout and inscribe zones applied for policy evaluation.
Includes the full frame with all transforms applied for policy evaluation.
Select which object types to detect and track.
frames
Run detection every N frames (default 5). Higher values decrease processing time but may miss fast-moving subjects.
px
Height in pixels to downscale frames for detection (default 1080). Set to 0 for native resolution.
High Low
How aggressively to detect objects (current threshold [[ Number(fancam.options.confidence_threshold).toFixed(2) ]], default 0.25). Higher sensitivity detects more but may include false positives.
Maximum number of objects to track per class, per clip (default 10).
Rows
Columns
Optional. Splits each frame into a rows-by-columns grid before detection so smaller or distant objects can be found more reliably. Leave both fields blank to use standard full-frame detection. [[ fancamSlicingError ]]
Videos submitted in this batch will be treated as one continuous recording — object tracking and crop videos carry across file boundaries. Submit video files in chronological order.
Policy Test must be performed in Async Mode in order to enable Multi-Event Processing.
If an event is flagged as noncompliant, confirm the assessment by assessing the current event together with the next event(s)
Number of events to assess together. Higher counts may provide more context but can increase latency.
Maximum time in seconds to wait for future events. This can help ensure timely responses even if the threshold of events is not met.
If an event is flagged as noncompliant, assess the combined current and future events against a different policy. Falls back to the current policy if the selected policy is not accessible.
Use this to OVERRIDE your prompt text.
Melting GPUs... ([[ elapsedTime ]] seconds)

Displaying [[ resultsData.length ]] Results

[[ getResultHeaderLabel(result) ]] [[ result.title ]]
Relevant Text Assessment Recommended Message
[[ policy['text'] ]] [[ policy['assessment'] ]] [[ policy['recommended_message'] ]]

[[ file.errorMsg ]]

Browser Support Tips:
  • Videos: Use MP4 (H.264/AAC) or WebM. H.265/HEVC is often not supported.
  • Images: JPG, PNG, WebP, and GIF are universally supported.
[[ file['origin'] || (resultFiles(result).length > 1 ? 'File ' + (index + 1) : 'Source') ]]

[[ result['title'] ]]

[[ formatJson(result) ]]
View in the Result Viewer 
[[ resultsJsonString ]]