Licensing
Live Caption Pro is commercial software. There is no public price list and no public download: access and the licence come with a quote.
In short
No public price, no public download. We answer every enquiry with a quote that fits the size of your environment.
© 2026 Jomaidon Technology Kft. All rights reserved.
What the licence covers
The caption application, the services behind it, the command-line tools and the web administration pages – everything we ship. The licence types and how many machines and users each one covers are set out in the quote.
What we say up front
- The voice service's current engine is Coqui XTTS-v2, whose licence (CPML) excludes commercial use. This applies to the cloned voice and to the service's 58 built-in voices alike – that is, to every voice mode of the outgoing direction.
- Before the commercial release we replace the engine; the candidates have been surveyed. Until then the voice service's voices may be used only in non-commercial, private trials.
- The in-process voice (Piper on the sherpa-onnx runtime, Apache 2.0; the licence of each voice model has to be checked per voice) is independent of this – but it lives only in the incoming direction.
- VB-CABLE is donationware and is not distributed with the product: you install it yourself from its maker's site.
- The model files are not part of the product. You download them under their own licences: the Whisper models, TranslateGemma, XTTS-v2 and the Piper voices each under their own terms.
Components in use
Open-source components run inside the product, each under its own licence. The first table is the list from the caption application's About window, including the model engines that run locally; the second is the four services'. The two overlap in part, because the services also run on your machines. Both are kept in step with the source repository by a test.
On your machine
| Component | Version | Licence | What we use it for |
|---|---|---|---|
| .NET | 10.0 | MIT | the runtime of every component |
| NAudio | 2.2.1 | MIT | audio capture and playback on Windows |
| Whisper.net | 1.9.1 | MIT | the .NET binding of the speech recogniser |
| whisper.cpp | MIT | speech recognition (Whisper) on the GPU or CPU | |
| llama.cpp | MIT | running the translation model locally | |
| sherpa-onnx | 1.13.5 | Apache 2.0 | running the in-process Piper voice |
| Piper | MIT | the in-process voice without a GPU | |
| Coqui XTTS-v2 | CPML | voice cloning in the voice service | |
| Avalonia | 12.1.3 | MIT | the user interface of the Avalonia client (the only one on macOS) |
| CommunityToolkit.Mvvm | 8.4.0 | MIT | the view models of the caption window |
| Serilog | 4.3.0 | Apache 2.0 | structured logging |
| Hardcodet.NotifyIcon.Wpf | 2.0.1 | CPOL 1.02 | the tray icon |
| OpenAI .NET | 2.8.0 | MIT | OpenAI-compatible client library for the translator's endpoint |
| Markdig | 1.3.2 | BSD 2-Clause | rendering the built-in help and this website's guides |
| Fluent UI System Icons | MIT | the icons of the macOS client (redrawn in place of the Windows icons) | |
| VB-CABLE | donationware | virtual audio cable between the call application and the interpreter | |
| FFmpeg | LGPL 2.1+ | decoding video and audio files for batch subtitling |
In the services
| Component | Version | Licence | What we use it for |
|---|---|---|---|
| ASP.NET Core | MIT | the web framework of the four services (.NET 10) | |
| MudBlazor | 9.5.0 | MIT | the component library of the web administration pages |
| Markdig | 1.3.2 | BSD 2-Clause | renders the help pages of the web administration |
| Serilog.AspNetCore | 10.0.0 | Apache 2.0 | logging in the services |
| Npgsql | 10.0.3 | PostgreSQL License | the account service's connection to its PostgreSQL database of users, roles and keys |
| PostgreSQL | 18 | PostgreSQL License | the database server of the account service – runs separately (in a Docker container or installed), the product does not ship it |
| Whisper.net | 1.9.1 | MIT | speech recognition in the caption server |
| Whisper.net.Runtime (whisper.cpp) | 1.9.1 | MIT | the native Whisper engine behind Whisper.net |
| llama.cpp | MIT | runs the translation model in the translator | |
| Coqui TTS | 0.27.5 | MPL 2.0 | the speech-synthesis library of the voice service (the XTTS-v2 model itself is under CPML) |
| Transformers | 4.57.6 | Apache 2.0 | model loading in the voice service |
| PyTorch | BSD 3-Clause | the tensor engine of the voice service, on the graphics card | |
| FastAPI | 0.111.0 | MIT | the HTTP interface of the voice engine |
| uvicorn | 0.30.0 | BSD 3-Clause | the web server of the voice engine |
Versions as pinned in the source repository's project and requirement files; an empty version means the installer takes the latest release.
Common questions
- Is there a subscription?
- Not for the processing: the models run on your own hardware, and there is no per-minute or per-call fee. The terms of the licence itself are in the quote.
- Where does my data go?
- Nowhere outside your network. The application runs on your Windows machine or Mac, the services on the same machine or on other machines you run. Nothing is sent to us.
- Can several machines share one graphics card?
- Yes. The services can run on one machine with a graphics card and serve several workstations on the network; the workstations then need no graphics card. How many machines and users a licence covers is part of the quote.
- Is the source code available?
- No. The source code is not public: the repository is private, and access to the product is given on request.