---
name: reels-lipsync-turbo
description: "Lip-sync on MiniMax H3 Max Turbo image-to-video — your own start frame (GPT Image 2.5 Flare 2K, vertical 1152×2048) plus the voice line in target_audio_url, one pass, no separate lip-sync model. Faster in the source test than H3 Max; the start frame fixes the look exactly. Part of the reels reference set."
---

These creative references describe source experiments with external tools. Use your own supported account and tools; this download installs no automation. Settings and timings are experimental observations, not current FlashBoards availability. Use only media and voices you have permission to use.

# reels-lipsync-turbo · start frame + voice on Turbo

> **Alternative external route:** Seedance 2.5 reference-to-video (`bytedance/seedance-2.5/reference-to-video`, references as @Image1… / @Audio1…, 4–30 s, native sound) was preferred in the source experiments. The measured side-by-side comparison was a product commercial, not every format; compare it with the H3 method below for your material.

Inputs are the person's own: their {look sheet} (or {photo}) for the start frame, {line} voiced with their {voice}.

1. **Start frame**: `openai/gpt-image-2.5/flare/edit`, quality low, `image_size 1152×2048`, {look sheet} as picture 1 — 27 s. Prompt structure used: "Picture 1 is a character sheet of one <man/woman/person>. A vertical 9:16 live-action film still of the same <man/woman/person>: <the place and light>. <framing>, looks straight into the lens, <mood>, about to speak. <Their> face, hair and clothes exactly as in picture 1. <film look>."
2. **Video**: `minimax/h3-max-turbo/image-to-video` · `image_url` = the frame · `target_audio_url` = the line (≥ 2 s) · `duration` = line length rounded up · `768P` · enhance disabled.
   ≈ 5 s generation for the 8 s source clip.

Prompt that worked:

```text
Live-action vertical shot, one continuous take from the given first image. The <man/woman/person> speaks the words of the pinned voice track straight to the camera, <their> lips exactly in sync with the voice: "<the line>". Small natural head movements and blinks while <they> talk(s), a faint smile on the last word. The camera holds still with a slight handheld breath. Everything else stays exactly as in the first image. No music, no other sounds.
```

Result: lips follow the words, blinks, the smile on the last word; the soundtrack is the given line (speech-to-text matches word for word).
Use Turbo when the exact frame matters (set, outfit, hair) or for many takes; H3 Max (`reels-talking-head`) for the most natural "creator" acting.

## Example run (reference only — never reuse its files, prompts, voices or images for a person)

The filenames below are historical output labels, not delivered or required files. In your project, use your own start frame, permissioned voice line, prompt and clip outputs through a supported external tool; no local runner is included.

- Demo character Mara Vale, 9.10.2026: start frame from Mara's sheet, a night laundromat (`try_ep1_start_a.webp` / `_b.webp`).
- Line "Rule number one: never go back for what you left behind. Rule number two... tomorrow." (`line_v4.mp3`); prompt `p_lipsync.txt`, job `job_lipsync.json`, clips `lipsync_v4_a.mp4`, `lipsync_v4_b.mp4`.