A small, open-source Mac dictation app that uses your ChatGPT sign-in. Speak into T3 Code, your editor, or any text field.
Contributing
Start an issue for a new feature or a change to sign-in or audio handling. Small fixes can go directly to a pull request.
Local setup
- Install Xcode Command Line Tools and Python 3 on macOS 13 or later.
- Clone this repository.
- Run
swift test. - Run
./scripts/build.shto createdist/GPT Voice.app.
Use ./scripts/build.sh --install to install your build in ~/Applications.
A local build uses ad hoc signing unless you select a certificate.
Changing its signature can require new microphone and Accessibility grants.
Use the same signing identity across local builds when possible.
Set GPT_VOICE_SIGNING_IDENTITY or put the identity in .local-signing-identity.
Do not commit certificates, keys, or credentials.
Checks
- Run
swift testafter a behavior change. - Check the actual app in light and dark mode after a UI change.
- Keep temporary capture files and account data out of tests and issues.
- Explain what changed, why, and how you checked it in your pull request.
The website is a static site in website/dist.
Serve it with python3 -m http.server 4173 --directory website/dist.
Its interactive preview uses scripted sample text. It does not access the microphone or ChatGPT.
Your contribution uses the project's MIT license. Third-party assets retain their own licenses.