Automated tagging
Generate category tags per video or per scene without manual catalog work.
Adult video metadata infrastructure
PHAR Adult Video Intelligence API
Viewers search, skip, rewatch, and leave once they find what they came for. PHAR turns long videos into navigable scenes and reusable metadata for search, recommendations, review, and editorial workflows.
- 20
- action categories
- Scene
- level metadata
- API
- or batch delivery
Input
Long-form videos, back-catalog batches, or new uploads.
Processing
Windowed action and position recognition with label smoothing.
Output
Tags, timestamps, chapters, annotated previews, JSON, CSV, or WebVTT.
Sample output
Metadata your video platform can actually use.
Turn adult videos into timestamped action and position metadata: tags, chapters, annotated previews, and candidate clips for review or editorial workflows.
Chapters
Consecutive windows are merged into stable scene ranges for timelines, player markers, search filters, and editorial review.
- 00:00 - 00:20Intro / Other
- 00:20 - 00:58Kissing
- 00:58 - 01:45Oral
- 01:50 - 04:00Doggystyle
Exports
Batch jobs and API calls can return exports that plug into CMS, search, moderation, or player systems.
{
"video": "sample.mp4",
"duration": "04:00",
"tags": ["Kissing", "Oral", "Doggystyle"],
"chapters": [
{"time": "00:20-00:58", "label": "Kissing", "confidence": "91%"},
{"time": "00:58-01:45", "label": "Oral", "confidence": "93%"},
{"time": "01:50-04:00", "label": "Doggystyle", "confidence": "88%"}
]
}Clip candidates
Candidate clips are ranked from detected categories, duration, label stability, confidence, and motion/activity signals.
- 00:24 - 00:42Kissing - 84% clip score
- 01:04 - 01:28Oral - 89% clip score
- 02:10 - 02:52Doggystyle - 86% clip score
Use cases
Built for catalog enrichment, search, recommendations, review, and clip workflows.
Timestamp generation
Let viewers jump directly to specific scenes instead of scrubbing through the full video.
Video chapters
Merge raw windows into stable sections for player timelines and navigation.
Search & recommendations
Use scene timing and action labels to improve related videos, discovery, and personalization.
Review cues
Route segments to human review based on category, score, or duration.
Shorts candidates
Rank segments for previews, teasers, and short-form clips using labels, duration, confidence, and motion.
Fast validation
Start with a batch pilot. No integration required.
The fastest way to test PHAR is to process a real sample from your catalog and inspect whether the metadata is useful.
Send a test batch
Share 10-100 videos that you are allowed to process with an external vendor.
Receive metadata
Get tags, timestamps, chapters, annotated previews, and structured exports.
Evaluate fit
Review quality on your catalog before discussing API, Docker, or licensing.
Commercial paths
Use PHAR as a service, API, or licensed model package.
Batch pilot
One-off processing for a selected catalog sample.
Best for validationHosted API
Send videos or URLs and receive timestamped metadata.
Best for recurring uploadsDocker / model license
Run the model in your own infrastructure with support.
Best for privacy-sensitive teamsBuyout discussion
Discuss broader model, code, and commercial-rights transfer.
Best for strategic buyersBoundaries
Clear about what PHAR is and is not.
- PHAR is adult action and position recognition, not age verification.
- PHAR does not determine consent, identity, or legal rights.
- Clip candidates are ranked suggestions, not guaranteed highlights.
- Validate fit on a sample batch before scaling to the full catalog.
Next step
Test PHAR on a real sample from your catalog.
A small batch pilot is enough to decide whether timestamped adult video metadata improves your catalog, search, or review workflow.