/v1/extractionsData extraction API
Pull the details you need from a post. Use a custom prompt to say what to look for and a schema to define what comes back.
Extraction docsSend a public post link. Get structured JSON or a timestamped video transcript.

Miso butter noodles in 15 minutes
Put timestamped transcripts behind a search box. Try a word from the sample clips and open the matching passage.
/v1/extractionsPull the details you need from a post. Use a custom prompt to say what to look for and a schema to define what comes back.
Extraction docs/v1/transcriptionsRetrieve existing captions or generate a transcript from the video’s audio. Choose plain text, sentence timestamps, or word timestamps.
Transcription docsWe’re building nData so apps and agents can get the data they need from different sources in a form they can use.
Launch targets are individual public posts on YouTube, Instagram, TikTok, LinkedIn, Reddit, X, Threads, Facebook, and Pinterest. Profiles, feeds, search results, and private sources are outside the current scope. Source availability depends on deployment settings, and a post may not have captions.
Configurable defaults allow videos up to 3 hours, up to 500 MB of aggregate selected media per post, and batches of up to 1,000 URLs. Credits, daily processing allowances, and other quotas still apply. These limits do not promise throughput, availability, or completion times.
The current default is a one-time grant of 100 starter credits for an eligible verified personal workspace. Extraction reserves 5 credits and transcription 2 credits per job; those rates are provisional. No automatic refill is active, and the future refill policy is still pending. Check your workspace usage for the effective balance, or contact support about credits or questions.
No. It uses static browser fixtures and never creates a job. Signed-in API and console submissions are separate and use the configured private processor. Read the API docs or request access if signup requires approval.
Paste a link from a supported platform to extract details or get a transcript.