Your editable transcript will appear here.
Processing happens on your device.
Review names and unclear speech before publishing.
One file. A useful first draft.Formats, privacy & limits ↓
A LITTLE CONTEXT
Use this Instagram transcript generator with an audio or video file you already have permission to process. Save the Instagram Reel or video yourself, then select that file here.
This page does not fetch or resolve Instagram links. Your Instagram video transcript comes from the audio in the file you choose, not the post description or on-screen text.
Copy the editable text or download TXT, Markdown, SRT or VTT. TXT and Markdown use your edited text; subtitle files keep the original timed segments. Speaker labels are not generated.
Up to 25 MB and 10 minutes per file. MP3, WAV, MP4, M4A, MOV, WEBM, OGG, FLAC and AAC are accepted; the audio codec must also be decodable by your browser. A shorter file is easier on devices with limited memory.
Whisper tiny is a small multilingual model. It can mishear words, especially names, accents, music and overlapping voices. Its language list is not a guarantee of equal accuracy across languages.
The first run loads model files from this site. Processing speed depends on your device. There is no paid transcription service or monthly quota.
GOOD QUESTIONS
No. Save the media yourself first, then choose the saved file. This page only handles files that you select or drop into the upload area.
Yes, if your saved file has audible speech. The model listens to the audio. Silent files cannot produce a transcript, and music-only or noisy audio may not produce useful text.
No. File decoding and speech recognition run in this browser. The site downloads its own model and runtime files; it does not upload your audio to a server. No account, history, analytics or third-party transcription service is used.
The browser must load and initialize the speech model and its runtime. The measured load time appears above. Later files can reuse the model while this tab stays open.
A supported file extension does not guarantee a supported audio codec. Try a WAV or MP3 version of media you own, or use a current browser. Files without an audio track cannot be transcribed.
Yes. Plain-text and timestamp views keep edits separately. TXT and Markdown download the currently visible text. SRT and VTT use the original timed segments to preserve subtitle timing.