Natural Speech Context — Concept
What I imagine with Natural Speech Context is the ability to start doing something in my codebase and simply talk out loud about what I am doing, with Windsurf using that speech as context for its autocomplete functions.
For example:
- I tab to a Python file
- Click into a function
- Start writing a new feature that requests a URL and saves the result
At that point, I could say:
“get URL variable and save result”
and Windsurf would pick up that instruction naturally.
Preferably:
- No keybind should be required to start recording.
- Or, optionally, a single toggle keybind could enable/disable listening.
- The user should be able to talk freely without constantly interacting with the keyboard to manage speech input.