Can I talk to a pull request?

Yes, on DiffGuardian Pro. On the pull request page, hold a key, ask what the change does, and hear the walkthrough spoken back while the dock scrolls to and highlights the lines it is explaining. It is a live spoken walkthrough: no setup per pull request, and you steer it by asking.

#How it works

1. Input voice and Output voice are both on by default for Pro. Turn either off in Settings → Voice.
2. Hold Space in the voice dock on the pull request page, ask, and release to send.
3. The dock answers out loud, scrolling to and highlighting the code as it explains it.
4. It can summarise the features in a pull request, walk you through one of them, or answer a question about the change. Ask a follow-up, or skip to the file you care about.

The dock has a Replay button that repeats its last answer. With Input voice on, the comment composer also gets a Dictate button, so a spoken comment is written up as clean text.

#Not the same as Trace It

Trace It is the written walkthrough, which you read at your own pace. Talking to a pull request is the spoken version, live and steerable. They are separate features. Talking to a pull request highlights the code it explains. It does not draw diagrams. The graph that builds while it talks belongs to talking to your codebase.

#Where speech goes

You choose a provider in Settings → Voice: your operating system's built-in voice, OpenAI or ElevenLabs. OpenAI and ElevenLabs use your own API key, and your voice goes straight from your machine to the provider you chose. The system voice needs no key. None of it goes to DiffGuardian's servers. See Voice for setup.

#Limits

• Talking to a pull request requires a Pro plan.
• OpenAI and ElevenLabs need your own key, entered in Settings → AI providers.
• Voice input and voice output can each be turned off in Settings → Voice.