YouTube URL or ID
Accept standard video URLs, Shorts URLs and normalized 11-character video IDs.
A YouTube transcript API built for production.
Send a YouTube URL or video ID and receive captions as plain text or structured segments. When captions are missing, the same endpoint can start a generated transcript job.
Transcript request
GET /v1/transcriptcurl 'https://capslane.com/v1/transcript?url=dQw4w9WgXcQ&mode=auto' \
-H 'x-api-key: YOUR_API_KEY'Native captions firstAI fallback when neededStable timestamped JSON
One endpoint for every transcript path.
The response contract stays predictable whether Capslane finds an existing caption track, serves a cached result or generates the transcript from audio.
Text or segments
Choose one plain text value or structured chunks with offsets, durations and language.
Language selection
Request a preferred ISO language and inspect every caption language found on the source.
Asynchronous generation
Long-running audio transcription returns a job ID with monotone progress and a polling endpoint.
Shared cache
Repeated requests reuse canonical transcripts while keeping tenant usage and logs isolated.
Operational visibility
Review request status, latency, cache use and failures from the workspace dashboard.
From video URL to usable text.
Validate
Capslane normalizes the YouTube identifier and rejects unsupported sources before making an upstream request.
Extract
The native path retrieves manual or automatic captions and selects the requested language when available.
Generate only when required
Auto mode starts local speech transcription only after a genuine absence of captions is confirmed.
Return one contract
Your application receives the same content fields for native, generated and cached transcripts.
Questions developers ask.
Does YouTube provide an official transcript endpoint?
The YouTube Data API does not expose a general endpoint for downloading transcripts from arbitrary public videos. Capslane provides a stable server API for that workflow.
Can the API return timestamps?
Yes. The segmented response includes text, offset and duration in milliseconds for every chunk.
What if a YouTube video has no captions?
Use auto mode to try available captions first and start an asynchronous generated transcript only when no usable track exists.
How many languages are supported?
Capslane returns the language used and the caption languages found on the video. You can request a preferred ISO language code.
Try the API on a real YouTube video.
Start free with no credit card and move to a paid plan when the integration grows.