FlashBoards user skills

Author bounded reusable local-board behavior with FlashBoards SDK 0.1 — Experimental, optional panels and live operation discovery.

Route the person's recurring AI character in stills — the character sheet from their photos, the same character across images and scenes, the character in a film frame — with GPT Image 2.5 edit at quality low, picture roles only, one pass.

Keep the person's AI character the same across many images — the character sheet as picture 1, every other reference with one job named by role and its opposite, the camera reference as the base frame, the light formula as a condition, in one GPT Image 2.5 pass; check with crops.

Put the person's AI character into a film frame in four modes — the whole character, the head only, the face only, the face keeping the frame's paint or make-up — with GPT Image 2.5 edit at quality low, blind role-only prompts and a look sentence that keeps the gaze off the lens.

Make the character sheet of the person's AI character from their photo(s) — face close-up and full body in one image, a turnaround, a back view, pose and expression sheets from reference charts — with GPT Image 2.5 edit at quality low, short role-only prompts and panel-by-panel checks.

Route animating the person's AI character with MiniMax H3 — a film scene from its frames, a cheap storyboard take on Turbo, a whole minute in shot pairs, motion from a video, a voice line, character + place + prop images, props that hold, prompt expansion — on a low-cost H3 route or the Seedance 2.5 quality route.

Make the person's character say a given line, or move to a given soundtrack — H3 Max Reference to Video with the sheet as Image 1 and the line as Audio 1 (the lips follow it), or Seedance 2.5 with @Audio1; on Turbo a given audio only plays under the clip.

Remake a film scene up to 15 s with the person's character in every shot — H3 Max Reference to Video with the character sheet as Image 1, up to eight frames of the scene (actor replaced) and the scene's audio, a shot-by-shot JSON prompt; Seedance 2.5 as the quality route.

Some finishing steps require an available editor or another integration.

• The film corpus gives frames, not the scene's sound: the audio has to come from the person, or the whole take goes on the Seedance quality route, which makes its own sound.
• The film corpus cannot list the frames of one scene: a film is browsed in shuffled order (about 1,500 frames, 24 a page), nothing filters by scene or timecode, and similar search returns other films only. Several shots of one scene come best from the person's own clip of it.

Animate one shot of the person's character from separate pictures — the sheet as Image 1, a location as Image 2, a prop as Image 3 — on H3 Max Reference to Video with prompt expansion off; small images padded, not scaled.

Choose H3 prompt expansion — disabled for any scripted scene, balanced or quality for a loose idea the model should pace, and read the expandedPrompt that comes back, because the video follows the rewrite, not your words.

Keep an unusual prop's shape in video — draw it on the character with GPT Image 2.5, build a start frame in the location, then animate from that frame on H3 Max Turbo with the action in three steps; never describe, count or forbid the prop in the video prompt.

Remake a film scene longer than 15 s with the person's character — pairs of shots on H3 Max Turbo from real start and end frames (actor replaced), a face board as the end image so the face holds, the board cut off and the pairs joined with the scene's sound.

Some finishing steps require an available editor or another integration.

• The film corpus gives frames, not the scene's sound: the audio has to come from the person.
• The film corpus cannot list the frames of one scene: a film is browsed in shuffled order (about 1,500 frames, 24 a page), nothing filters by scene or timecode, and similar search returns other films only. Several shots of one scene come best from the person's own clip of it.

Animate a film scene up to 15 s cheaply on H3 Max Turbo — the scene's frames (actor replaced) as one 3×3 storyboard first image, the character sheet as the last image, both far off the timeline, the scene's sound laid under; trim the one-frame flashes.

Some finishing steps require an available editor or another integration.

• The film corpus gives frames, not the scene's sound: the audio has to come from the person, or the whole take goes on the Seedance quality route, which makes its own sound.
• The film corpus cannot list the frames of one scene: a film is browsed in shuffled order (about 1,500 frames, 24 a page), nothing filters by scene or timecode, and similar search returns other films only. Several shots of one scene come best from the person's own clip of it.

Make the person's character do the motion of a clip with the clip's camera — H3 Max Reference to Video with the sheet as Image 1 and a short silent clip as Video 1, conformed to exactly 16:9 or 9:16, 24 fps and 2, 5, 10 or 15 s; for Turbo, the motion in words.

Some finishing steps require an available editor or another integration.

• Only the person's own video results can be conformed on the server; a clip from elsewhere must already be 16:9 or 9:16, 24 fps and 2, 5, 10 or 15 s to go in as a public URL.

Route vertical Reels, Shorts and TikTok with the person's recurring AI character talking to camera — voice, emotion, lip-synced or prompt-spoken video, room sound — on a low-cost H3 route or the Seedance 2.5 quality route.

Put emotion and delivery into the character's line with ElevenLabs v4 audio tags ([whispers], [laughs], [sighs], free descriptions), punctuation for pauses and emphasis, and stability; check that no tag is read aloud.

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.

Make a generated talking head sound recorded in a real room — a short room reverb on the voice and an ambience mixed about 13 dB under it, generated with ElevenLabs Sound Effects or taken from a real recording; or re-shoot on Seedance 2.5, which makes its own synced sound.

Make the person's AI character talk to camera like a real creator's story — H3 Max Reference to Video with the look sheet as Image 1 and the voiced line as Audio 1, 9:16, the blind prompt that works; Seedance 2.5 as the quality route.

Choose and keep the voice of the person's AI character for a Reels or Shorts series — ElevenLabs v4 with a saved clone, a natural public-library voice or a designed voice; voice every line with it and check it.

Clone the person's voice from their own consented recording (or from a designed voice), keep it as a saved FlashBoards voice and reuse it for every line; check, replace or delete it.

Route every upscale — close-up, full figure, character sheet, film still, start frame, a cheap re-draw, skin, a finished video, video takes — to the right Magnific or Topaz setting, scale chosen by the input size and the price quoted first.

Enhance the skin of a big face only — Magnific Skin Creative (pores and freckles pushed, face kept), Faithful (subtle) or Flexible (the most detailed) — after or instead of the sublime upscale; never on a full figure or a whole sheet.

Upscale a face close-up or portrait with Magnific Precision V2 sublime — 4× on a small 1K close-up, 2× with a grainy analog look on a bigger frame — after checking the face is right; skin as an optional second pass.

Upscale a film still, screenshot or small reference with Magnific Precision V2 sublime — 2× with a grainy analog look by default, 4× only when the size is needed; it cleans noise and blocks, keeps faces and the lens look, invents nothing.

Upscale a full figure, a wide shot or small faces with Magnific Precision V2 sublime — 4× on a standalone 1K figure, 2× with a grainy analog look on a bigger frame or half a sheet — never with skin enhancement.

Re-photograph the person's own draft frame at higher resolution with GPT Image 2.5 and a strict prompt when they ask for it by name — cheap and fast but not pixel-exact; every real upscale is Magnific.

Upscale a character sheet best by upscaling the close-up and the full figure as separate images (Precision V2 sublime 4× each while a half is about 700 px, 2× on bigger halves), or cheaply in one 2× pass; never skin enhancement on the whole sheet.

Make the start frame of a video at the size the detail must survive — GPT Image 2.5 Flare 2K for a close-up, 4K for a wide shot, or a 1K frame upscaled with Magnific — because the video keeps the detail drawn into the frame, not the file size.

Enhance the finished, edited video once with Topaz Starlight Precise 2.6 (to 4K or 1080p) through enhance_video on the person's own cut, quoted first and confirmed; trim before, lay the original sound back after.

For takes in progress, generate at the final resolution instead of upscaling — H3 Max Turbo 768P for drafts, 1080P for final takes — from a well-drawn start frame, with reference images far off the timeline, then enhance the edited cut once.

Route an AI music video on a whole track starring the person's character — one film as the world, a linear three-act story, every face replaced, H3 Max Reference to Video in parts of up to 15 s with eight shots, transitions on the beat and lip-sync — on a low-cost H3 route or the Seedance 2.5 quality route.

Some finishing steps require an available editor or another integration.

• MCP reference search returns previews, not full HD masters, exact corpus scene units or a scene_data export; use an uploaded scene or supplied shot data.
• Arbitrary crop/mask/feathered paste is not exposed by the current media-tool catalog; use a suitable discovered image-edit workflow or upload the repaired frame.

Assemble the music video — trim every part to its planned length, join the parts, lay the person's original track as the only sound, check the cuts against the plan, then enhance the finished cut once.

Choose one film whose story fits the track, write a three-act story with the person, cast the roles, and find frames for every story beat in the film corpus — four frames a part, the lead alone where possible.

Some finishing steps require an available editor or another integration.

• The corpus is searched by what a frame shows and browsed by film; a search by script line or dialogue is not available.
• A film is browsed in shuffled order (about 1,500 frames, 24 a page), not by scene or timecode, and text and similar search cannot be limited to that film; narrow a browse with the interior, exterior or night filter instead of paging through the whole film.

Replace every face in the music video's frames — the lead becomes the person's character (with their look sheet, or the film's costume with the face board), every other person gets an invented face — with GPT Image 2.5 edit, one person per pass, every face checked with crops.

Some finishing steps require an available editor or another integration.

• Arbitrary crop/mask/feathered paste is not exposed by the current media-tool catalog; use a suitable discovered image-edit workflow or upload the repaired frame.

Write and generate every part of the music video — H3 Max Reference to Video with the face board, the look sheet, four frames and the part of the track as Audio 1, a JSON prompt of eight shots with cuts on beats, a different transition each, light, acting, head movement and lip-sync.

Read the person's track for a music video — lyrics with word times, the style in words, the measured beat grid, and parts of up to 15 s cut on a beat at the longest vocal pause — and cut the audio of every part.

Route remaking a film scene with the person's own characters — the scene's cuts, setups and lines read from the clip, one replaced picture per camera setup, the actors' own performance re-voiced in the characters' voices with the scene's room tone, H3 in parts of up to 15 s that end on a cut — on a low-cost H3 route or the Seedance 2.5 quality route.

Some finishing steps require an available editor or another integration.

• The film corpus gives frames, not the scene's clip, sound or subtitles: the person brings the clip of the scene (an upload) and the rest is read from it.

Read the scene to remake as data from its clip — the cut times, the camera setups, the lines with their times and speakers, the parts of up to 15 s that end on a cut, and the acting shot by shot.

Some finishing steps require an available editor or another integration.

• The corpus has no scene clips, subtitles or shot lists here: the person uploads the clip of the scene, and cuts, lines and speakers are read from it.
• A scene cannot be looked up by its number or timecode: text search covers every film and a film is browsed in shuffled order, so the person names the scene and brings its clip.

Make one replaced picture per camera setup of the scene — the person's characters in the main roles, every other face invented, no text or watermark — with GPT Image 2.5 edit and blind role-only prompts; check every face with crops.

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.

Re-voice the scene's dialogue in the person's characters' saved voices — isolate the voices, change each speaker span into the voice of the character who plays that role, lay the spans back at their times on a track exactly as long as the generation, and add the film's own room tone.

Route AI advertising for a product (or the person's own) in three formats from one product sheet — a UGC video with the recurring character, a 15 s commercial, static creatives with the offer text — on a low-cost H3 route or the Seedance 2.5 quality route; ElevenLabs for voice and music only.

Make the product sheet every ad image and video takes as its reference — four views, a big label close-up and the hero state, every word of the label in quotes — with GPT Image 2.5 Flare at quality medium for crisp text.

Make a 15 s product commercial — three keyframes with the product (macro hero, the action, the end card with the slogan), a music bed with a drop, six shots on the beat — as an H3 Max Reference to Video draft or a Seedance 2.5 final with its own synced sound effects.

Make static ad creatives — a 1:1 feed post, a 9:16 story with the character, a 16:9 banner — with GPT Image 2.5 Flare at quality medium drawing the product, the scene and every line of the offer text exactly as written.

Make a UGC ad of up to 15 s — the person's character confesses, opens and tries the product and ends on a call to action — with the voice as Audio 1 on H3 Max Reference to Video (the lips follow it), or the lines in a Turbo prompt and the voice changed after; Seedance 2.5 as the quality route.

Route an AI avatar that speaks a voice track of any length from one picture and the voice — a standing talking head, a solo video podcast or YouTube host in 16:9 with two angles, or a walk-and-talk room tour — on a MiniMax H3 Max Turbo stub + Extend chain, with Seedance 2.5 as the quality route.

Make a talking clip longer than one model call — the voice cut at pauses into pieces of up to 14 s, a silent 2 s stub from the person's picture on H3 Max Turbo, one H3 Max Turbo Extend step per piece with output continuation, assembled with free trims and joins and the stub cut off.

Make a solo video podcast or a YouTube host in 16:9 — a breakout episode found with vidIQ for the structure, the person's own script in segments of about 45 s, two angles of one studio on GPT Image 2.5 Flare, one H3 Max Turbo stub + Extend chain per segment with the angles alternating, joined with the clips' own sound, and a thumbnail.

Some finishing steps require an available editor or another integration.

• A reference episode's transcript is not available through vidIQ here: take its structure from the title, length and what the research returns.
• vidIQ thumbnail generation is not available here: draw the thumbnail with the image models; titles and their scores come from youtube-packaging.

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.

Route a YouTube channel with a recurring AI character — what to make from vidIQ research (breakout videos, trends, keyword demand), a reference episode's structure, the person's own script, the host on the H3 Max Turbo stub + Extend chain with two angles, and the title, description, chapters and thumbnail ready to paste.

Some finishing steps require an available editor or another integration.

• Channel numbers are public data: private YouTube analytics (retention, traffic sources, audience, revenue, thumbnail click-through rate) stay in YouTube Studio.
• Uploading to YouTube needs social_publish on this client (social); otherwise the person uploads the file and pastes the packet.

What works on a YouTube channel and why, from its public numbers — growth over a window, its most popular and latest videos with breakout scores, how fast its videos usually gather views — read as three findings and three actions; vidIQ calls paid per call, private Studio analytics stay in YouTube Studio.

Some finishing steps require an available editor or another integration.

• Public channel numbers only: retention, traffic sources, audience, revenue and thumbnail click-through rate stay in YouTube Studio.

Titles and a thumbnail for a YouTube video with scores — scored title ideas from vidIQ from a working title or a published video, a score for the person's own title, a thumbnail drawn from the look sheet, and vidIQ's score for the thumbnail of a published video; scores are predictions, the real test is YouTube Studio's Test & Compare.

Some finishing steps require an available editor or another integration.

• A thumbnail is scored only once it is on YouTube; vidIQ thumbnail generation is not available here, the thumbnail is drawn with the image models.

Route a short animated cartoon with the person's recurring character in any style — one 3×3 storyboard picture on GPT Image 2.5 Flare, one 15 s H3 Max Turbo image-to-video request whose voices, foley and music come from the model, the board flash cut off; Seedance 2.5 as the quality route.

Draw the 3×3 storyboard of a cartoon episode in one GPT Image 2.5 Flare request at 2048 × 1152 — the character sheet as picture 1, the style in words, nine frames with their actions, numbers in the corners, black gutters, no speech bubbles — and check the grid before the video.

Generate a 15 s cartoon episode from one 3×3 storyboard on H3 Max Turbo image-to-video with the model's own voices and sound — the board as the first frame only, 480P, 15 s, prompt expansion off, no end frame, no target audio, the prompt that keeps the board off screen — then cut the board flash.

Route making a song with Suno V6 — from the person's hum, sung or rapped line (the melody is kept), from an idea, from their own lyrics, in the style of a reference they name (MusicBrainz), or as a remake of their own track; then stems, a longer version, fixed words, cover art and a music video. One choice per turn.

What to do with a finished take — pick a take, stems (vocal + instrumental by default, multi-stem only on request), a longer version with Suno V6 Audio Extend, fixed words or style as a new round, square cover art, and a music video.

Turn the person's hum, sung or rapped line into a full song on Suno V6 that keeps their melody and flow — the words with Scribe and the kind, genres, tempo and theme with Audio Understanding, three choices (words, vibe, lyrics), lyrics written around their line, a Suno V6 cover of the seed.

Turn an artist, a song or a genre the person names into music directions from a MusicBrainz snapshot — the artist's genres, tags, era and similar artists, the best-known artists of a genre and era, a song's genres, and a words-only style line for Suno without names. Free and read-only.

Write song lyrics and the Suno style line for the person — structure tags, the hook from their own words, the language of the seed, length by duration, the style formula (genre + voice + instruments + tempo + mood, no artist names) and negative tags; also a song from an idea or from their own lyrics.

Route "what is working and will mine work" for TikTok, Instagram Reels and YouTube Shorts — breakouts in a niche with the hook of the first 3 seconds and the format (vidIQ), a draft checked by measured facts and an eight-criterion rubric into a band per platform with three fixes, the platforms' rules for length, AI labels, reach and money with source and date, and a publish pack for the finished video; then the picked format goes to the mode that makes it.

Some finishing steps require an available editor or another integration.

• A post measured against its account's usual and a scene-by-scene of a post are not available yet; what works in a niche, a draft check and the platform rules are.

What is working right now in a niche on TikTok, Instagram Reels and YouTube Shorts — one vidIQ search for TikTok and Reels and one for Shorts return posts that did many times better than their account's usual, each with the hook of the first 3 seconds, the format, the pacing and the sound where vidIQ reads them; then at most three picks, a pattern line and the FlashBoards mode that makes each format.

Check a draft vertical video before posting on TikTok, Instagram Reels or YouTube Shorts — measure it with FlashBoards tools (length, size, cuts, first word, words in the first 3 seconds), lay out two frame sheets, score eight rubric criteria from the sheets and the facts, and get a band per platform with the platform checks and three fixes from get_draft_verdict; also compares two versions. A band reads the craft and is not a forecast of views.

Some finishing steps require an available editor or another integration.

• Loudness, silence, black and frozen frames are not measured on the server: sound and pacing are scored from the words, the cuts and the frame sheets.
• The score log that checks bands against real results is not available yet, so every band is unchecked.

The platforms' rules that decide reach and money for a short video, each with its source and date, from get_platform_rules — length and format thresholds, ranking signals the platforms named, AI labels and what happens without them, originality and re-upload rules, monetization conditions and benchmark numbers with their formula, for TikTok, Instagram Reels, YouTube Shorts and long YouTube, X, Snapchat, Pinterest, LinkedIn, Threads and Facebook.

Package a finished video for each platform before the person posts it themselves — the versions (9:16 for TikTok, Reels and Shorts; 16:9 for YouTube) with conform_video or a crop, the platform checks and AI policies from get_platform_rules, a caption and hashtags in the person's own words, and a one-screen list of what goes where; then, where this client offers social_publish, each version is drafted for the person's approval (social), otherwise the person uploads it.

Some finishing steps require an available editor or another integration.

• Posting needs social_publish on this client (social); otherwise the person uploads each version.
• Video versions are 9:16 or 16:9 only.

Post the person's finished video to their own connected social accounts with their approval — connect YouTube, TikTok, Instagram, Facebook, Threads or X (through FlashBoards' apps or its partner Zernio), draft each version with the captions, hashtags and AI label from the publish pack, hand over the approval page or in-chat card, follow each network until the post is live, cancel what has not gone out, and read comments and counters where the network supports it. Only where this client offers the FlashBoards Social tools; otherwise the person uploads each version.

Some finishing steps require an available editor or another integration.

• FlashBoards Social is open to a limited audience: where social_connect and social_publish are not among this client's tools, the person uploads each version (viral-publish).
• Comments and counters come from the networks that report them, through FlashBoards' own YouTube app or Zernio; an account that cannot be read is named. Answering comments is not available yet.
• Every post needs the owner's approval on its page or in-chat card; nothing is posted from the chat alone.