AI
Video & Audio to Text
Turn a recording into text you can search, edit and reuse.
Drop video or audio file here, or click to upload
Supports MP4, WebM, MOV, MP3, WAV, M4A and more
File limit: 500 MiB
Transcription mode
Fastest and lightest — Whisper tiny
Best for mobile devices and slower connections
Better accuracy — Whisper base
Better for desktop and clearer recordings
Audio language
Choosing the language manually gives faster, more stable results.
Runs locally in-browser
Your file is processed on your device and is not uploaded to our server.
First run caches the AI model in-browser.
Use Cheetah for phones and tablets. Dolphin suits desktops with more memory.
Processing speed depends on source length and your hardware.
Local CPU processing. Long recordings take more time; try Cheetah or a shorter sample.
Make the spoken content easier to work with
Interview notes, lecture summaries and subtitle drafts start with an accurate record of what was said. ClipGG runs speech recognition on your device and shows editable segments. Select the spoken language, listen back to uncertain passages and correct names before treating the transcript as a quotation.
How to use it
Open a file, select its spoken language and a model, then transcribe and review the segments.
Practical example
For an interview, test a short recording before processing the full file. Keep a list of names and specialist terms beside the transcript. Choose SRT when you need timed captions; choose text when the next task is reading, searching or preparing notes.
What to know before you start
The limit is 500 MiB. Models download on first use. Accuracy and speed depend on speech, hardware and browser support.
Questions about this tool
Should I choose Cheetah or Dolphin?
Cheetah uses the smaller model and suits limited memory. Dolphin uses a larger model; compare a short sample on your device before choosing it for a long recording.
Why does the first run take longer?
The recognition model must be downloaded before processing. Connection speed affects that first step; recording length and hardware affect the transcription itself.
Will SRT export give me finished subtitles?
It provides text with timing, but names, punctuation and segment boundaries still need review. No automatic transcript is guaranteed accurate enough to publish unchanged.