Short answer
That error string matches the Speech-to-Text âCreate transcriptâ endpointâs hard limit for direct file uploads: the file field in a multipart/form-data request must be under 3.0 GB (â3072 MB). If youâre instead sending cloud_storage_url, that path has a lower 2 GB limit. (elevenlabs.io)
What to check
Confirm which product/endpoint youâre calling and how youâre supplying the media:
Dubbing in the web app: 500 MB and 45 minutes per clip; Dubbing API allows up to 1 GB and ~2.5 hours. (elevenlabs.io)
Voice Isolator in the web app: up to 500 MB or 1 hour. (help.elevenlabs.io)
Voice Changer (voice conversion) input duration: up to 5 minutes. (help.elevenlabs.io)
If your file is âmuch shorterâ but you still see 3072 MB:
Make sure youâre using the expected request shape: the STT endpoint âexpects a multipart formâ and you must provide exactly one of file or cloud_storage_url. Mixing both can cause validation errors. Use the API reference âTry itâ to verify the same file succeeds. (elevenlabs.io)
If you share which endpoint you called, whether you used file or cloud_storage_url, and the exact file size on disk, I can advise precisely against the documented limits above. (elevenlabs.io)
raw canopy
tiny crow