Training Your Assistant by Voice¶
What it does¶
Voice is the primary and most natural way to train your assistant. Tap the microphone, speak freely about yourself, and warpSpeed auto-transcribes and sends your message when you go quiet. The AI responds and asks follow-up questions, it's a real conversation.
How to use it¶
Starting a voice session¶
- Open the AI Training screen.
- On the entry screen, tap the "Just talk freely" card (mic icon, bottom of the screen). This immediately puts the app into listening mode.
- Alternatively, if you've already had a conversation, tap the floating mic button (green pill button) at the bottom of the chat thread.
The listening screen¶
Once listening starts, the app switches to the Listening screen: - A warpSpeed ring animation plays (animated ring GIF with the warpSpeed logo in the centre). - The heading reads: "I'm listening, [your first name]" - Below the animation: "Press the button when you're done" - A large red Stop button is shown at the bottom.
Speaking¶
- Speak naturally. You don't need to structure your sentences, just talk about yourself, your preferences, your goals, your schedule, your habits, or anything that might help warpSpeed understand you.
- Auto-send on silence: when you stop speaking for a moment, warpSpeed automatically transcribes and sends your message. You don't need to tap Stop unless you want to end early.
- The heading changes to "I'm ready, [your first name]" once transcription is ready and the keyboard/input is shown.
After speaking¶
- Your transcribed message appears in the conversation thread as a user bubble.
- warpSpeed processes your message and replies, often asking follow-up questions to learn more.
- Tap the mic button again to record another voice message in the same conversation.
Stopping the session¶
- Tap the red Stop button to end voice input immediately and return to the chat view.
- Your transcribed message is sent automatically unless you stop before speaking.
Tips & shortcuts¶
- Speak in British English for best transcription accuracy; the speech-to-text engine is set to the
en-GBlocale. - Don't worry about filler words ("um", "like"); the AI filters those out during processing.
- After sending, check Warp's knowledge of me to see what the AI extracted from what you said.