Natural introductionWarm, clear, comfortable pace
RequiredCloned narration walkthrough
See the complete voice-to-video flow.
This local simulation lets you try the interface when ElevenLabs is not configured. It demonstrates what is saved, what stays exact, and when MiniMax receives the approved audio.
Demo onlyNo recording is uploaded, no clone is created, no provider is contacted, and no credits are charged.
- 1Capture the voice
- 2Generate and time script
- 3Send approved audio
01
Not createdVoice clone
In Assets, give the voice a name and record at least one clean sample.
Expressive deliveryCuriosity, excitement, and range
OptionalNumbers and detailsDates, prices, and reference codes
OptionalThe real button securely uploads permitted recordings to ElevenLabs for zero-shot cloning.
02
Waiting for voiceNarration studio
The spoken script remains separate from the visual video brief.
Timing check
Not generatedListen before generating
- Estimated voice
- —
- Video length
- 6s
Playback uses your browser voice for demonstration only. The real workflow synthesizes the exact script with your saved ElevenLabs clone.
03
LockedApproved audio handoff
The preview you approve is the same audio file sent to MiniMax H3.
Controls the scene, camera, performance, pacing, and product presentation.
Creates one measured MP3. Build Prompt may align visuals but cannot rewrite these words.
Receives the approved MP3 bytes for speech and visible performance synchronization.
reference_audio: waiting for an approved previewNot sentDemo handoff complete. In production, CraftyAI now submits the approved MP3 as MiniMax H3 reference audio. The existing video price is debited only after successful video generation.