Metadata API Coming soon

Pull structured data out of a master — subtitles, taxonomy tags, and performer identification. Three endpoints under one API, each callable on its own against content you host.

Coming soon — not callable yet

None of these endpoints are live. The page documents the intended shape so you can build against it ahead of release; calls will 404 until the Metadata API ships. Today, transcription and tagging run inside the pipeline (POST /v1/scenes).

Runs standalone

Point any endpoint at a master you host and get structured data back — no need to move your catalog into Jetwild.

Transcription

Speech-to-text with word-level timing. Returns .srt / .vtt subtitle tracks in 30+ languages for SEO, accessibility, and on-site search.

POST/v1/transcribeGenerate subtitle tracks

POST/v1/transcribe

Request

{
"file_url": ".../8841.mp4",
"languages": ["en", "es"],
"formats": ["vtt"]
}

Response

Press Run transcribe to see the response

vtt / srt for players and tubes; json for word-level timestamps. Pass languages to fan one master into several localized tracks.

Tagging

Auto-tag against your own taxonomy — categories, acts, attributes, and niche routing — so a years-deep back-catalog becomes searchable and merchandisable.

POST/v1/tagGenerate taxonomy tags

POST/v1/tag

Request

{
"file_url": ".../8841.mp4",
"taxonomy": "spizoo_v3"
}

Response

Press Run tagging to see the response

Map tags to your category tree, not a generic one. Register a taxonomy once and every scene routes into the vocabulary your site already uses.

Model ID

Identify performers and match them to your roster, returning your internal model IDs. Unmatched faces come back flagged for review, never guessed.

Identification, not verification

Model ID never asserts age or consent. Those come from §2257 records — and §2257 is only required when you distribute. See §2257 & data security.

POST/v1/models/identifyDetect + match performers

POST/v1/models/identify

Request

{
"file_url": ".../8841.mp4",
"roster": "spizoo"
}

Response

Press Run identify to see the response

Bulk & async

Every endpoint has a /batch variant (up to 500 masters) and an async mode — pass a webhook_url or subscribe to scene.ready for results as they finish.

Pricing

Each endpoint meters per minute of video, with a discounted bulk tier for back-catalog runs. See Pricing & metering.