Grok Bot can talk now.
Turn a spoken request into a visible work product. Give voice, processing, and document separate states.
View HTML source
<!doctype html><html lang="zh-CN"><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><title>Grok Bot can talk now. · sequence</title><style>html,body{margin:0;width:100%;height:100%;overflow:hidden;background:#fffaff}canvas{display:block;width:100%;height:100%;object-fit:contain}</style><body data-mode="sequence"><canvas width="1000" height="600"></canvas><script src="data.js"></script><script src="common.js"></script><script src="scene.js"></script><script src="bridge.js"></script></body></html>Motion and composition
Local motion · Composition rules
Anchor the voice overlay to the phone
Place a white chat phone in a purple-orange glow, aligning microphone, waveform, and chat content along one vertical center. Use green and purple to distinguish voice visuals and clear sans serif text.
After speech ends, enter the working state
View HTML source
<!doctype html><html lang="zh-CN"><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><title>Grok Bot can talk now. · sequence</title><style>html,body{margin:0;width:100%;height:100%;overflow:hidden;background:#fffaff}canvas{display:block;width:100%;height:100%;object-fit:contain}</style><body data-mode="clip" data-start="2" data-end="24"><canvas width="1000" height="600"></canvas><script src="data.js"></script><script src="common.js"></script><script src="scene.js"></script><script src="bridge.js"></script></body></html>Expand the waveform after the microphone is clicked; when speech ends, settle the waveform and switch to processing. Restore voice visuals for another response, avoiding the same looping waveform throughout.
Resolve the request into a document card
Introduce a clear text task later, then expand its corresponding document. Keep document name, version, and task topic related, giving the voice-triggered change a concrete outcome.
Replace the task and conversation
Adapt to reports, schedules, or design deliverables. Determine the user's speech, processing state, and final output first, then coordinate waveform and subtitle timing.
Sections and rhythm
Click a section to seek to it in the choreography example above.






Visual references · Discrete still frames
Creating other themes
- When adapting to another assistant, preserve speech→processing→visible output.
- Supply sound and its actual waveform separately; the example waveform is illustrative only.
Production constraints
Phone chat / microphone / voice overlay / waveform / working state / document card
- Attach the voice overlay to the phone, then move from the response to a specific task and document.
- Waveform, processing, and result are different states; do not display the same waveform continuously.
- The generated document name must match the request topic.