Transcribe multiple files at once
Download SnailText, point it at a folder, and it transcribes every recording inside in one run — one file at a time, with per-file progress so you can see exactly where the batch is. A month of call recordings, a stack of interviews, a term of lectures: all done in a single pass.
A folder can mix audio and video. Audio formats run directly; video runs after a one-time one-click decoder download. Everything happens on your own machine.
A transcript generator for a whole folder
Think of it as a transcript generator that works offline — an audio and video transcript generator in one — files in, transcripts out, no upload and no per-minute charge no matter how big the folder is. A hundred hour-long recordings costs the same as one — nothing — because your own hardware does the work.
Whether you call it batch or bulk transcription, the flow is the same: point it at the folder, walk away, come back to a set of transcripts. That also keeps confidential material where it belongs — when the folder is full of customer calls or NDA-bound interviews, nothing is handed to a server.
Separate files or one combined document
When the batch finishes, you choose the layout. Export separate files — call_01.mp3 becomes call_01.txt next to it — or a single document with each recording under its own heading, named by the source file.
Formats are text, Markdown, or JSON. JSON is handy when the transcripts feed into another tool or script.
Why batch beats transcribing one file at a time
Transcribing recordings one by one means babysitting the process: open a file, wait, save, open the next. Batch transcription is set-and-forget — you queue the whole folder once and it works through everything unattended, so a backlog of interviews or a month of call recordings clears in a single pass instead of a dozen manual rounds.
It is the right tool when you have a pile of files rather than a single recording. If a file in the batch is unsupported or unreadable, SnailText skips it and keeps going, so one bad file does not stop the run. For a single recording, audio to text or video to text is the simpler path.
Runs on any computer — a graphics card is optional
Transcription itself does not need a GPU. SnailText runs the Whisper speech model on your CPU, so batch transcription works on an ordinary laptop with no graphics card — it is just faster when a GPU is present.
The optional AI post-processing step is the part that wants a GPU. Turn it on and a local language model cleans up each transcript in the same run — fixing errors, adjusting the writing style, or translating the text into another language. That step is a Pro feature and needs a graphics card with 6 GB of VRAM or more. Without it, you still get the full, unlimited transcription on any machine.