--- name: avatar-walk-talk description: Make the person's avatar walk while talking — a room tour, a show-around, a walk and talk — on the stub + Extend chain: a selfie start frame with the whole place visible, the route split by the voice pieces (one stop per piece, gestures tied to words), the handheld-selfie piece prompt. --- # Walking avatar and room tour 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 `avatar` and `avatar-long-speech` for the shared rules. Inputs are the person's own: {look sheet} or {picture}, {speech} with their {voice}, {place}, {route}. The walk stays one chain (at most 60 s): the place holds only inside one clip. ## Steps 1. **Start frame** — every stop of {route} visible behind the character, so Extend keeps one place: `generate_image` with `gpt-image-25-flare`, `quality: "low"`, `width: 1152, height: 2048`, `referenceGenerationIds: []`: ```text Picture 1 is a character sheet of one . A vertical 9:16 live-action selfie frame shot with a phone's front camera held at arm's length; the phone itself is never seen. The same stands of <{place}>, the place visible behind : . A medium close-up from chest up, slightly high angle, look(s) straight into the lens with a smile, about to speak. face, hair and clothes exactly as in picture 1. Natural phone camera look, . ``` 2. **{speech}** — written along {route}: one stop about every 10 s, a line about each thing. Then cut it into pieces as in `avatar-long-speech` and **assign the stops to the pieces by their transcribed words** (`speechChunks`), not by the plan. 3. **Piece prompt** — for every piece: ```text A vertical 9:16 selfie : the films with the phone's front camera held at arm's length in one hand; the phone itself is never seen. Audio 1 is voice: say(s) exactly these words, lips exactly in sync with Audio 1: "" walk(s) and what do(es) on which word>. The camera moves with as walk(s), the framing bobs and sways a little with steps like a real handheld phone, and the place passes behind . face, hair and clothes stay exactly as in the video, and it is the same as in the video: . . One continuous shot, no cuts, no music, no text on screen. ``` - `` shape: " walk(s) over to , the camera following face, and on '' tap(s) / point(s) at / hold(s) ." — one physical action per thing, a laugh, sigh or whisper where the speech has it. - Repeat the same list of things in every piece: the only memory of the place Extend has is the clip so far. 4. **The chain** — the stub, one Extend step per piece with `output: "continuation"`, the trims and joins, the stub cut off (`avatar-long-speech`). ## What was measured - About a minute of speech in 5 pieces of 10–12 s gave one 60 s clip in a few minutes. - The speech is less faithful than with a standing speaker: steps and room sound get mixed in or the words re-voiced, and the clip drifts up to about 3 s. **Listen to the whole clip before publishing.** ## Check `transcribe_media` the clip against the script; `extract_video_frame` at each stop and `describe_media` "Is it the same person and the same place as the start frame? What do they point at?" Ask also "How many people are in the frame?": a returned Extend piece once held a second figure (a shoulder in the foreground and a ghost trail) without any error. Retake that piece once with " is the only person in the place, always facing the lens" in its prompt, and re-join the pieces after it. ## 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-talking-avatar/skills/avatar-walk-talk.md) - Updated from the owner's skill sources of 2026-10-11. Read: `avatar`, `avatar-long-speech`, `reels-voice`. Source experiments and demo materials are historical; use the person's own approved media.