- file
- local-transcription.transcript
- speakers
- You, vocateca
- address
- vocateca.com/en/local-transcription
- duration
- 00:02:57 · 398 words
Transcribe your own files, fully offline
Opening
Drop an audio or video file into Vocateca and get a timestamped transcript on your Mac. Nothing is uploaded — the file never leaves your machine.
How it works
Drop a file. Drag in any common audio or video file — meetings, lectures, voice memos, interviews. Batches are fine. Transcribe on-device. The model runs on your Mac. There is no upload, no server, and it keeps working without a network connection. Read and export. Read the timestamped transcript, correct names if you like, and export to Markdown, SRT, VTT, or JSON.
How it's different
Cloud transcribers upload your recording and bill per minute. Vocateca runs entirely on your Mac — private, offline, and free to run as often as you like.
| Feature | Vocateca | Online tools |
|---|---|---|
| Where audio is processed | On your Mac, offline | Uploaded to their servers |
| Works without internet | Yes | No |
| Cost model | One app, no per-minute fees | Per-minute or subscription credits |
| Batch files | Drop many at once | Often one at a time |
| Sensitive recordings | Never leave your machine | Stored on their servers |
| Export formats | Markdown, TXT, SRT, VTT, JSON | Limited or paywalled |
| CLI + MCP | Yes — scriptable, LLM-ready | No |
| Speaker labels | Optional diarization | Often paywalled |
In short. If a recording is private, keep it that way — transcribe it on your own machine.
What leaves your Mac for this source
Your files themselves never leave your Mac. From the network, the app fetches only the speech models and ffmpeg for them, once, the first time they are needed. The entries that belong to no single source, such as updates, are on the complete list too. The complete list
| # | What | Goes to | Only when |
|---|---|---|---|
| The first time a feature needs it2 entriesonce | |||
| 01 | Speech models | The model weights for Parakeet-TDT and Whisper, and for Qwen3-ASR if you choose it, plus the models that tell speakers apart. No network needed afterwards.huggingface.co | Parakeet-TDT while the app sets itself up. Whisper and Qwen3-ASR the first time you use them. |
| 02 | ffmpeg, ffprobe | A pinned release, checked against a fixed SHA-256 hash before it runs.github.com | While the app sets itself up, unless ffmpeg is already on your Mac. |
More to explore
Every export format. Markdown for notes, SRT/VTT for captions, JSON for pipelines. Pick what your workflow needs. See export formats
Works offline. No network required. Transcribe on a plane, in a SCIF, or anywhere your recordings shouldn't touch the cloud.
Private by default. Transcription runs on-device. No uploads, no third-party servers, no account required. How privacy works
Transcribe from the command line
The same engine ships as a CLI and an MCP server, so you can transcribe a local file from a script or hand it to an LLM agent.
Terminal
vocateca transcript ./interview.mp3 --format mdOpen-core: the CLI and MCP server are Apache-2.0.
End of transcript
Transcribe your files, privately