Kits AI's Transcription service converts your audio recordings into structured text using a high-precision AI engine. Meetings, interviews, voice notes, podcasts — everything becomes a downloadable text document in minutes.
Get started for free →Drag and drop your audio file into the designated area, or click to browse from your device. MP3, M4A, WAV and WMA formats are accepted, up to 100 MB.
Recordings with a clear voice and no background noise give the best results.
Click "Send file". The audio is first uploaded to our servers, then passed to the AI transcription engine. Processing is entirely asynchronous.
You can close the tab once the upload is complete — transcription continues in the background.
Find your task in the Transcription list. The card shows the real-time status: pending, processing, or completed. The status updates automatically.
A 10-minute audio recording typically takes 1 to 3 minutes to transcribe.
Once processing is complete, the card switches to "Completed" status. Download the transcription as a text file from the card menu. The file contains the full timestamped text.
The text file is ready to copy, edit or integrate into a document.
MP3
Universal format, compatible with all devices. Ideal for dictaphone recordings or podcasts.
M4A
Apple format, produced by iPhones and Macs during voice recordings.
WAV
Uncompressed, high-quality format. Larger files but no audio signal loss.
WMA
Windows Media Audio format, produced by Windows devices and some dictaphones.
🎙️
The clearer the audio, the more accurate the transcription. Use a dedicated microphone and record in a quiet room.
🗣️
The transcription faithfully captures all speech. For multi-speaker meetings, make sure voices do not overlap too much.
📁
The limit is 100 MB. For long meetings, split the audio into 30 to 60-minute segments for faster processing.
📝
AI transcription is very accurate but may stumble on proper nouns and highly technical terms. A proofreading pass is recommended before publishing.
MP3, M4A, WAV and WMA, up to 100 MB. If your file exceeds this limit, split it into segments using a tool like Audacity or GarageBand.
The engine automatically detects the spoken language and transcribes accordingly. French is the primary optimized language, but English and other common languages also work well.
The cost depends on the audio duration. It is shown before launching — you will never be charged without confirmation.
As a general rule, expect about 10 to 20% of the audio duration. A 10-minute recording is processed in 1 to 3 minutes.
Currently, the transcription is available as a text file (.txt). Other formats (Word, SRT for subtitles) are planned for future versions.
Uploaded audio files are stored securely on Google Cloud Storage. You can delete a transcription at any time from the list.
Welcome credits included — no credit card required.
Get started for free →