Practical overview
When it fits
Use it when a Unity application needs local Whisper transcription and the target platform is among those listed in the repository documentation.
Who it is for
- Unity developers adding speech recognition or transcription to an application.
- Teams building local voice-command input, speech-to-text features or audio-clip transcription.
- Developers who need a Unity implementation spanning desktop and listed mobile platforms.
Documented uses
- Transcribe speech within a Unity application without relying on an internet connection.
- Add voice-command recognition to a game or other Unity experience.
- Turn recorded audio clips into text.
- Choose model weights to balance transcription speed and quality.
Requirements & limits
The repository is public and licensed under MIT. The README says the project can be used in commercial projects.
- The included ggml-tiny.bin model is small and fast but is described as lower quality than other models.
- Model sizes offer speed and accuracy trade-offs; English-only weights are cited as an option for improving English transcription.
- The README lists Windows, macOS, Linux, iOS, Android and visionOS support, while marking WebGL unsupported.
- Vulkan acceleration is listed for Windows and Linux; Metal for macOS, iOS and visionOS. Metal requires Apple7 GPUs or newer, beginning with M1.
- The project includes prebuilt native libraries; rebuilding them requires the documented whisper.cpp version and platform-specific process.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Add whisper.unity to a Unity project
Bring the binding and its included examples and tiny multilingual model into a Unity project.
Before you start
- A Unity project and Unity Package Manager.
- Internet access to retrieve the repository package.
Steps
Open the repository as a regular Unity project, or add it through Unity Package Manager.
To add it as a package, use https://github.com/Macoron/whisper.unity.git?path=/Packages/com.whisper.unity.
Use the included examples and tiny multilingual model weights as the starting point.
Expected outcome
The Unity project has the whisper.unity package or repository, plus the included examples and small model weights.
Guide limitations
- The README describes WebGL as unsupported.
- The bundled tiny model is described as lower quality than other model options.
Select and add different Whisper model weights
Replace or supplement the bundled model with another weight file.
Before you start
- A whisper.unity Unity project.
- A model weight file compatible with whisper.cpp.
Steps
Choose model weights from the whisper.cpp model collection.
Place the downloaded weights in the Unity project's StreamingAssets folder.
Select model size or language-specific weights according to the speed and accuracy trade-offs described by the project.
Expected outcome
The alternative model weights are located in StreamingAssets for use in the project.
Guide limitations
- The README does not prescribe a particular model for a given device or accuracy target.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 14 October 2026
- Original website title
- Macoron/whisper.unity · Official homepage title · retained in its source language