---
name: reels-talking-head
description: "An AI character talking to camera like a real creator's story — MiniMax H3 Max reference-to-video with the character sheet as Image 1 and the voice line as Audio 1, 9:16, 768P; the Instagram prompt that worked, what broke (a phone in frame, the outfit swapped) and the blind fix. Tested at 8 s. 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-talking-head · H3 Max, sheet + voice

> **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: Image 1 = their {look sheet} (or {photo}), Audio 1 = {line} voiced with their {voice}.

`minimax/h3-max/reference-to-video` · `reference_image_urls: [<look sheet>]` · `reference_audio_urls: [<line audio>]` · `duration: 8` · `resolution: 768P` · `aspect_ratio: 9:16` · `prompt_expansion_mode: disabled`.
7.5 s generation · ~13 s request. Output 768 × 1344.

## The prompt (blind — nothing about the sheet is described)

Fill the slots; keep the sentence order (it is what was measured). Use the character's pronouns.

```text
Image 1 is only the look of the <man/woman/person>. <They> wear(s) exactly the outfit of Image 1 and <have/has> exactly the hairstyle of Image 1 — the same garments, colours and hair, nothing added, removed or swapped. Image 1 is a reference and never appears on screen. Audio 1 is <their> voice: <they> say(s) exactly these words, <their> lips in sync with Audio 1: "<the line>". A vertical 9:16 Instagram talking-head video shot with a phone's front camera; the phone itself is never seen. <They> <sit/stand> in <the place and light — 3–4 concrete background details, soft in the background>. <They> talk(s) straight into the phone camera, casual and natural, like recording a story for <their> followers. <acting beats tied to words>. The framing is slightly crooked and wobbles a little, like a handheld phone. Light: <the light sources of the place: glow, bokeh, flare, dust in the beam>. One continuous shot, no cuts, no music, no text on screen.
```

- `<the place and light>` example value: "a small bedroom in the evening: a warm desk lamp, fairy lights on the wall behind, an unmade bed and a plant soft in the background".
- `<acting beats tied to words>`: one beat per word or pause of {line} ("on '<word>' <they> lean(s) in closer to the lens", "pauses before '<word>'", a smirk, a laugh, a small gesture), matched to {mood}.
- Never describe the face, hair or clothes in words — only "exactly … of Image 1".

## What we learned (5 takes)

| Prompt | Result |
|---|---|
| "filmed on a phone propped up on the desk" | ✕ **the phone is in the frame** |
| "face, hair and clothes" from Image 1 | ✕ a casual black T-shirt, hair down — the "story at home" won |
| + "exactly the outfit and hairstyle of Image 1, nothing swapped" | ✓ the outfit back; hair still down |
| + acting beats tied to words (lean in on a key word, bite the lip before the last word, smirk, laugh, tuck hair) | ✓ every beat played, hair pulled back like the sheet |
| + light: lamp glow, bokeh, flare, dust | ✓ warm glow and bokeh |
| + "reverb of the room, a car passing" | ✕ **no ambience added** (silence level unchanged) → `reels-room-sound` |

Audio 1 is not played back 1:1 here (correlation with the input 0.23–0.64): listen before you publish.

## Example run (reference only — never reuse its files, prompts or images for a person; select a permissioned voice)

The Hope identifier below is an external library-voice sample, not private credentials or the required voice. Choose your own permissioned voice and reference images. The named audio and clips are historical output labels, not delivered or required files.

- Demo character Mara Vale, «Mara's Rules» ep. 1, 9.10.2026: Image 1 = Mara's sheet, Audio 1 = "Rule number one: never go back for what you left behind. Rule number two... tomorrow." in Hope (`uYXf8XasLslADfZ2MB4u`, `line_hope_v4.mp3`).
- Place: her small bedroom in the evening, warm desk lamp, fairy lights, unmade bed, plant. Beats: a small breath before she starts; leans in and narrows her eyes on "never"; bites her lip before "tomorrow", playful smirk and a soft breathy laugh; tucks a strand of hair behind her ear.