---
name: filmscene-shots
description: Write and generate one part of the remade scene — a JSON prompt with speakers, the setup images, every shot's cut, camera, action, acting and dialogue at its measured word times, off-screen lines marked — on H3 Max Reference to Video or Turbo, then check cuts, sync and faces before the next part.
---
# One part, one generation

FlashBoards user skill, version 1.2.3 (2026-10-13).

On the board, read these guides with `list_skills` and `get_skill` when the assistant exposes them; otherwise use the hosted index through an available read-only retrieval tool. Use that runtime’s actual model discovery, current page context and guarded board actions for execution. Keep native readiness, approval and cost gates; do not claim a guide was read when retrieval is unavailable.

Read related guides with `get_skill({skillId:"<id>"})` or the [hosted index](https://pub-78da82d51ed14147ac554d38094b0c0b.r2.dev/user-skills/1.2.3/index.html). Reading a guide does not start or authorize a paid generation.

Read `filmscene` for the shared rules.

> **Quality route:** Seedance 2.5 (`seedance-2-5`) with the setups as @Image1… and the voice track as @Audio1 does this job much better; quote both.

## Prompt (fill the slots from the person's cast)

```
speakers: {character A}: S1, {character B}: S2
{ output: one continuous film scene, D seconds, 21:9, N shots with hard cuts exactly at the given times; Audio 1 is exactly D seconds…,
  references: { Image 1: setup A — every {character A} shot looks exactly like it…, Image 2: setup B — every {character B} shot…, Audio 1: the whole dialogue… },
  characters: { {character A}: {look: the person facing the camera in Image 1…, voice: {character A}'s voice of Audio 1 (name it by what tells the two voices apart, e.g. the man's / the woman's / the lower voice)…}, {character B}: {…} },
  environment: { location…, light…, sound: only Audio 1; no music },
  style: live-action 35 mm film, natural and restrained…,
  shots: [ { shot, cut_at, from, to, image, camera (setup + "static"), action: what each person does from the start of the shot to its end, acting,
             dialogue: [ { speaker: S1, on_screen: true, when: 00:02.140–00:03.200, line: "…" } ] } ],
  rules: [only the person facing the camera speaks…; an off-screen line is heard from the person seen from behind and the face on screen keeps its mouth closed; no cut to the speaker inside that shot; one full-screen shot at a time; no phones] }
```

- Looks come from the setup images only (blind) — never write a character's face or clothes into `look`.
- Times are **the voice track's measured word times** (`transcribe_media` of Audio 1), not subtitle times.
- An off-screen line: `on_screen: false` plus the rule that the listener's mouth stays closed. H3 Max kept the listener's mouth closed and the line in place.
- Keep headings and service words away from the quoted lines: the model may speak words that stand next to a quote.

## Requests

- **H3 Max Reference to Video**: `generate_video` with `minimax-h3-max-reference-to-video`, `referenceGenerationIds`: [setup A, setup B], `referenceAudioUrls`: [the voice track's `audioFinishingTaskId`], `duration` D, `resolution: "480P"`, `aspectRatio: "21:9"`.
- **H3 Max Turbo**: `minimax-h3-max-turbo`, `firstFrameUrl` = the first setup, `endFrameUrl` = the last setup, `targetAudioUrl` = the voice track, the same D and 480P, `promptExpansion: "disabled"`; the prompt says "the given first frame / last frame / the given audio" instead of Image/Audio numbers. On Turbo the audio plays as given but the lips do not follow it — for lips on screen put the lines into the prompt and change the voices after, or use H3 Max Reference to Video.

## Checks after every generation

- Cuts: `detect_video_cuts` → the cut times against the scene's.
- Sync per line: `transcribe_media` of the generation → the word times against Audio 1.
- Faces and the off-screen moment: one frame a second (`extract_video_frame`), `arrange_images` into a sheet, `describe_media` with a question.
- Send the part to the person next to the original and wait for approval before the next one.

## Execute through MCP or on the board

Use the person's FlashBoards account; never ask for a personal provider key or run the source's local scripts. Discover `get_account`, `list_generation_models` and the tools of this client first: a published guide does not prove a feature is deployed.

**Headless, no open board.** Quote with `estimate_generation_cost` (skill models: `estimate_skill_model`), then run `generate_image`, `generate_video`, `generate_audio` or `run_skill_model` with one stable `generationAttemptId` per logical attempt. Pass `maxCostUsd` only when the current operation and selected model support it; otherwise use the quoted cost and the person's explicit approval with the supported controls. Never silently ignore an unsupported spending cap; make one `get_generation` check after `retryAfterSeconds`. Pass the person's own results by `generationId`: images in `referenceGenerationIds`, `firstFrameUrl` and `endFrameUrl`, audio in `referenceAudioUrls` or `targetAudioUrl`, videos in `referenceVideoGenerationIds`. They are read on the server; signed result links do not work as public URLs. Every `generationId` a tool returns stays valid: a later read of the same result gives a different string, and both work. Film frames and gallery pictures go in by `previewHandle`: `referencePreviewHandles` of the generation tools, a `{previewHandle}` source of `arrange_images`, `crop_image` and `pad_image`, and `describe_media`. New files come from `show_media_upload_picker` as receipts. The person's character comes in as its sheet, by `generationId` or as an upload; there is no search by name. If the person names a character without the sheet, ask for it or make one with `character-sheet`; do not scan recent results for it. Free server steps: `trim_video`, `concat_videos`, `conform_video`, `replace_video_audio`, `extract_video_frame`, `extract_video_audio`, `detect_video_cuts`, `crop_image`, `pad_image`, `arrange_images`, `cut_audio_span`, `concat_audio_tracks`, `mix_audio_tracks`, `fit_audio_duration`, `add_audio_reverb`. Check every result with `describe_media` (a question about an image; `model`, `mode` and `sourceRange` only for a video) and `transcribe_media` for speech. Both checks are paid: `describe_media` costs a fraction of a cent for an image and about 2–4 ¢ for a short video (more for a long one), `transcribe_media` under 1 ¢ a minute — count the checks you plan in the price you name before the first paid step. `describe_media` is not reliable about sound: it misses effects mixed under music and can hear voices or instruments that are not there. Check speech with `transcribe_media` and ask the person to listen for effects and music. Never repay an uncertain submission; reuse an attempt ID only for the identical request. Make each new `requestId` or `generationAttemptId` unique, for example the task, the date and a short random suffix: an ID used before, even in another chat, replays that earlier request or answers `REQUEST_ID_CONFLICT`.

**On the board.** Inspect with `list_board_sessions` and `get_board_snapshot` and carry the same `sessionId` and a fresh `activeBoardId` as `expectedBoardId`. Native cards: `execute_board_action` (`listNativeModels`, `setModel`), `prepare_board_generation` with ordered board items or owned outputs, then one `generate_on_board` within the authorized budget. Board media tools: `list_media_tools` → `start_media_tool` → `get_media_tool_run`. Put headless results on a board with `add_media_to_board`.

Keep mandatory safety settings and the live limits of each model. Prices, timings, endpoints and "free retry" claims in the sources are historical, not FlashBoards contracts. If a step this guide needs is not available or is refused (a tool error, `audio_not_configured`, a missing model or input), keep the finished assets, tell the person which step is blocked and why, and ask how to go on. A single generation that ends `failed` is not a blocked step: submit it once more with a new `generationAttemptId` (reworded if it may have hit the content checker), and stop to ask only if it fails again. A failed attempt is not always refunded, so the retry may be paid again and `maxCostUsd` caps each request on its own: count both against the budget the person approved. Never swap in another route on your own: no image model redrawing a layout or a frame, no `describe_media` in place of the frames it should check, no clip generated only to take its sound, no picture drawn where the guide takes a film frame.

## Source and related guides

- [Original Learn guide](https://flashboards-h01.vercel.app/learn/formats-and-exhibition/ai-film-scene/skills/filmscene-shots.md)
- Updated from the owner's skill sources of 2026-10-11.

Read: `filmscene`, `filmscene-corpus`, `filmscene-sound`. Source experiments and demo materials are historical; use the person's own approved media.