--- name: reels-lipsync-turbo description: Make a talking clip on MiniMax H3 Max Turbo from the person's own start frame — Turbo speaks the line from the prompt, then the speech is changed into the character's saved voice; targetAudioUrl only lays a track under the clip and does not move the lips. --- # Exact start frame and the line on Turbo 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:""})` 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 `reels` for the shared rules. Inputs are the person's own: their {look sheet} (or {photo}) for the start frame, {line} as text, their saved {voice}. > **On Turbo the lips do not follow a given audio.** `targetAudioUrl` only lays a track under the clip. So on Turbo the line is spoken from the prompt and the character's voice is changed in afterwards. When the lips must follow a given voice line, use `reels-talking-head` (H3 Max Reference to Video). > **Quality route:** Seedance 2.5 (`seedance-2-5`) with the sheet as @Image1 and the line as @Audio1 does this job much better; the method below is the low-cost route — quote both. ## Steps 1. **Start frame**: `generate_image` with `gpt-image-25-flare`, `quality: "low"`, `width: 1152`, `height: 2048`, `referenceGenerationIds: []`. Prompt structure: "Picture 1 is a character sheet of one . A vertical 9:16 live-action film still of the same : . , looks straight into the lens, , about to speak. face, hair and clothes exactly as in picture 1. ." If detail is short, `upscale-start-frame`. 2. **Video**: `generate_video` with `minimax-h3-max-turbo`, `firstFrameUrl: `, the line written in the prompt, `duration` = the line at about 2.5 words a second rounded up, `resolution: "480P"` (`"768P"` for a sharper take), `promptExpansion: "disabled"`. No `targetAudioUrl` for the voice. 3. **The character's voice**: `extract_video_audio` from the clip → `change_voice_audio` with the saved `voiceId` and `audioSource: {source: {audioFinishingTaskId: }, startSeconds: 0, endSeconds: }` → `replace_video_audio` with the clip and the changed speech. The timing and the lips stay; the voice becomes the character's. To keep background sound apart from the voice, first split the clip's audio with the Demucs stem model (`run_skill_model`), change only the vocals, then `mix_audio_tracks` them back with the other stems. Prompt that worked: ```text Live-action vertical shot, one continuous take from the given first image. The says exactly these words straight to the camera: "". Small natural head movements and blinks while 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. No subtitles, no captions, no text on screen. ``` ## When Turbo, when H3 Max Use Turbo when the exact frame matters (set, outfit, hair) or for many takes; H3 Max Reference to Video (`reels-talking-head`) for the most natural creator acting and whenever the lips must follow a given voice line or recording. `targetAudioUrl` on Turbo is only for a track or recording that must play under the clip as it is. ## Check - `transcribe_media` the finished clip: the words of {line}, nothing added. - `extract_video_frame` at the middle and `describe_media` with a question: captions burnt in? the same face and outfit as the sheet? ## 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-reels-and-shorts/skills/reels-lipsync-turbo.md) - [Original Learn guide](https://flashboards-h01.vercel.app/learn/formats-and-exhibition/ai-talking-avatar/skills/reels-lipsync-turbo.md) - Updated from the owner's skill sources of 2026-10-11. Read: `reels`, `reels-voice`, `reels-talking-head`, `upscale-start-frame`. Source experiments and demo materials are historical; use the person's own approved media.