---
name: reels
description: "Reference guide for vertical Reels / Shorts / TikTok with a recurring AI character that talks to camera — voice (ElevenLabs v4, voice library, voice design, clone), emotion tags, lip-synced video on MiniMax H3 (Max reference-to-video or Turbo), room sound. Classifies the job and the priority (fast / draft / best) and hands off to the reels-* sub-skill with exact settings, prompts and times. Use when the user says reels, shorts, tiktok, talking head, story, vertical video, my character speaks to camera, рилс, шортс, сторис, говорящая голова. Two external routes: the H3 draft method of this skill, and the quality route — Seedance 2.5 reference-to-video; the source preferred Seedance for quality, but the side-by-side measurement was a product commercial, not every format."
---

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 · reference workflow

> **Source author's sound preference (9.10.2026), not FlashBoards policy:** avoid ElevenLabs Sound Effects (`/v1/sound-generation`, `fal-ai/elevenlabs/sound-effects`) for effects, room tone, ambience and jingles; the author found them strange and louder than the voices and preferred ElevenLabs for voices (TTS v4, Speech-to-Speech, Audio Isolation). Historical demonstrations below and in related guides still include ElevenLabs Music beds and an earlier Sound Effects ambience test; those are separate experiments, not a universal voices-only rule. For your own mix, use permissioned film sound (line-free gaps, Demucs without vocals), the video model's native sound or supported video-to-audio tools; keep effects under the voice.

**Work for the person's own character: their photo / look sheet / voice from their character card. The example run below belongs to a demo character.**

Measured on source experiments (9.10.2026, see the example run). Article with the source examples:
https://flashboards.yaroflasher.com/learn/formats-and-exhibition/ai-reels-and-shorts/

Slots used in all reels-* skills: {character} · {photo} · {look sheet} · {voice} · {line} · {place} · {mood}.

## 🏆 Source experiment pick — one talking-head clip (8 s, 9:16)

1. **{line}** → ElevenLabs **v4** (`eleven_v4`) with {voice}: the character's saved voice_id / clone, or a natural library voice picked for them → `reels-voice`.
2. **Video** → **MiniMax H3 Max reference-to-video**: Image 1 = {look sheet} (or {photo}), Audio 1 = the line, 9:16, 768P, enhance disabled, the Instagram prompt with {place} and {mood} → `reels-talking-head`. 7.5 s generation.
3. **Room sound** → your own or licensed ambience of {place}, following the source preference above, mixed under the voice + a short room reverb in ffmpeg → `reels-room-sound`. That guide separately records the earlier ElevenLabs ambience experiment.



## Two external routes: draft or quality

| Route | Model | When |
|---|---|---|
| **Separate draft option** — not the measured 768P method above | MiniMax H3 Max / H3 Max Turbo, 480P | drafts and many takes; not the measured talking-head setup |
| 🏆 **Quality** | **Seedance 2.5** reference-to-video (`bytedance/seedance-2.5/reference-to-video`, `task: reference`): the same pictures as `image_urls` (@Image1…), the same audio as `audio_urls` (@Audio1…), 4–30 s a call, native synced sound, 480p / 720p / 1080p | the final cut, when the look matters |

**Source observation (9.10.2026): Seedance 2.5 was preferred for quality; H3 remains the draft route.** Measured side by side on a 15 s product commercial (`ad-spot`): Seedance 2.5 at 720p — cut deviations up to about 1 s in the listed measurements (versus up to 2 s on H3), finer liquids and detail, its own sound effects in sync — took 7 min against 9 s on H3 Max 480P. Not yet measured on every format; write the same shot-by-shot prompt in plain text and refer to the references as @Image1, @Audio1. Compare the routes before choosing a generation.

## Classify

| The job | Go to |
|---|---|
| Pick or make the character's voice | `reels-voice` |
| Clone a voice from a recording, keep it, reuse it | `reels-voice-clone` |
| Make the delivery emotional (whisper, laugh, sigh, pauses) | `reels-emotion-tags` |
| The character talks to camera like a real creator | `reels-talking-head` (H3 Max) |
| The exact start frame matters, or faster | `reels-lipsync-turbo` (H3 Turbo) |
| It sounds too clean, "AI" | `reels-room-sound` |

## Global rules

1. **The voice is fixed for the series.** Choose it once (ElevenLabs `voice_id`, clone id, or a Qwen speaker embedding file) and keep it in the person's character card / memory; voice every line with it.
2. **Prompts are blind**: never describe the reference images ("exactly the outfit and hairstyle of Image 1", not the garments or hair in words). Exception (owner, 9.10.2026): the expression of a reference frame — gaze, brows, mouth — may be named.
3. **H3 does not add ambience on request.** Room tone, street, reverb — mix them yourself after (`reels-room-sound`).
4. **Never mention a phone in the frame** ("filmed on a phone on the desk" draws the phone). Say "shot with a phone's front camera; the phone itself is never seen".
5. ElevenLabs Creator plan: **max 5 concurrent requests**; 30 voice slots for your own clones (library voices do not take slots).
6. Never substitute the demo's files, prompts, clone or images for your own material. Library voices listed as samples may be used only if you deliberately select them and have permission through your own account; their example identifiers are not private credentials or a default cast.

## External tools

Use supported ElevenLabs voice tools and the named video models through your own account. Check every audio with `fal-ai/elevenlabs/speech-to-text/scribe-v2`: the words, and that tags were not read aloud.

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

- Who: demo character **Mara Vale**, the series «Mara's Rules», episode 1 · 9.10.2026.
- Line: "Rule number one: never go back for what you left behind. Rule number two... tomorrow." (85 characters).
- Voice: library *Hope — Bubbly, Gossipy and Girly* (`uYXf8XasLslADfZ2MB4u`); also a designed-then-cloned voice `mara` (`iV8DaEoP6jYNZUgTCz0e`).

These are external sample settings, not delivered audio or a supplied clone. Select and save your own permissioned voice; no example files are required.