Instagram Reel to text: paste a link, transcribe on your device
Paste the link here and the next step opens InstaScribe with it already filled in. The speech model runs on your machine, not ours.
Step 1 — InstaScribe takes it from here
Or go straight to InstaScribe and paste the link there.
The numbers, and where each one came from
| Spec | Value | Where this comes from |
|---|---|---|
| Where the model runs | your own deviceWeights are fetched once and cached; the audio itself is never sent anywhere. | MeasuredshotMod |
| Audio it works from | AAC LC, 48 kHz, stereoThe track Meta's Reels specification describes, decoded to mono for the model. | OfficialMeta for Developers |
| Longest input | bounded by the Reel itselfMeta caps a Reel at 90 seconds, which keeps a transcript comfortably inside a phone's memory. | OfficialMeta for Developers |
| Text output formats | TXT, SRT, VTT, JSONProduced by InstaScribe once the transcript exists. | MeasuredshotMod |
| Accuracy on accented speech | varies with the model chosenNobody, including us, can promise a number for this — it depends on audio quality, accent and background noise. | UnverifiedshotMod |
Last verified 2026-09-05 · 4 sources
- Reels publishing — recommended video specification (9 x 16, 1080 × 1920, 3–90 s, 24–60 fps, H.264/H.265, AAC LC 48 kHz) — Meta for Developers
- Behaviour of the shipped in-browser export path, measured on the shotMod build dated 2026-09-05 — the tool doing it is the one mounted above — shotMod
- Derived arithmetically from the ratio and width in the cited Meta specification — no independent claim is made — shotMod
- Community consensus, not an Instagram document — Instagram publishes no current page stating this value, and help.instagram.com serves article bodies only to JavaScript, so no Help Center article can honestly be cited for it (checked 2026-09-05) — shotMod
What this actually means
Transcribing a Reel is the one job on this site where the interesting part is where the computation happens. Almost every transcription service uploads your audio to a server, runs a model there and sends text back — which means the audio lands on somebody else's disk. The alternative is to move the model to the audio instead, and modern browsers are fast enough to make that practical.
So this page is deliberately a doorstep rather than a workshop. It explains what the operation is and what it costs, and then hands your link to InstaScribe, where the speech model actually runs. The link travels in the address bar and the audio never leaves your machine, which is the entire point of doing it this way.
What goes wrong
The first run is slow and the reason is honest: the speech model has to be downloaded before it can be used. That is tens of megabytes for the default model and several hundred for the most accurate one, cached afterwards, and on a phone over mobile data it is a real cost. Doing the first transcript on a connection you are not paying by the megabyte for is a genuinely good idea.
A transcript is also not a caption file, and the gap between the two is wider than it looks. Running speech to text gives you words and rough timings; it does not give you line lengths that read well on screen, speaker labels, or punctuation you would be happy to publish. Expect to edit. If subtitles are the goal rather than a searchable text, start from the SRT page instead, which is built around cues rather than paragraphs.
How to do it here, in three steps
- Put your public Reel link in the box near the top and press Open in InstaScribe.
- InstaScribe opens with that link already loaded — this page mounts no transcriber of its own.
- Pick a model there, run it, and copy the text or export TXT when it finishes.
Questions, answered
How do I get the text of an Instagram Reel?
Paste the public Reel link on this page and continue to InstaScribe, which opens with your link already filled in. The speech model runs inside your browser, so the audio is never uploaded.
Is there a free Instagram transcript tool with no sign-up?
It is free and there is no account. Because the model runs on your own device rather than on a rented server, there is no per-minute cost to pass on to you.
Why is the first transcript slow?
The speech model has to download before it can run. That happens once and is then cached by your browser, so the second Reel starts almost immediately.
Public links only — shotMod resolves posts that are already public and never asks for your Instagram login. Only download or convert content you have the rights to.