add_audio
ChatGPTPut a voiceover, music or sound effects on a video. Mix it with the sound the video already has, or replace that sound. Set the volume of each, start the new sound later, loop a short music bed and fade it out at the end. The video keeps its length. Free. Returns a job id at once; call show_media with it, or get_job when the job is done.
add_captions
ChatGPTBurn captions into a video from its own speech. Every word is timed to the voice. Six looks, from large capitals that light up the spoken word to plain sentences on a dark box, the way short-form video does. Works in most languages. Costs 5 cents. Returns a job id at once; call show_media with it, or get_job when the job is done.
add_logo
ChatGPTPlace a picture on a video: a logo in a corner, a product shot, a sticker or an end card. Set its size, its place and when it shows. A PNG with a transparent background keeps its shape. Free. Returns a job id at once; call show_media with it, or get_job when the job is done.
approve_review
ChatGPTStart a run that waits for review, because the person pressed Generate on the Visual Sandbox card. Only the card calls this tool, with the token it holds for that run. Charges this account the price the card shows, and returns the job that started.
assemble_film
ChatGPTJoin a film's shots into the first cut, in shot order, and lay its sound. audio_mode narration replaces the shots' sound with the film's narration; mix lays the narration over it; shots keeps it. A music bed loops under everything and drops while the narration speaks, and the last frame holds when the narration runs past the last shot. Every shot needs a finished clip. Free. Returns a job id at once; call show_media with it, or get_film_project once it is done.
audio_eleven_music
ChatGPTRun Eleven Music, and charge this account for it. Eleven Music is ElevenLabs' text-to-music model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_elevenlabs_sound_fx
ChatGPTRun ElevenLabs Sound FX, and charge this account for it. ElevenLabs Sound FX is ElevenLabs' text-to-sound-effects model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_elevenlabs_tts
ChatGPTRun ElevenLabs Text to Speech, and charge this account for it. ElevenLabs Text to Speech is ElevenLabs' AI voice model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_elevenlabs_voice_changer
ChatGPTRun ElevenLabs Voice Changer, and charge this account for it. ElevenLabs Voice Changer is ElevenLabs' model for speaking a recording again in another voice. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_elevenlabs_voice_clone
ChatGPTRun ElevenLabs Voice Clone, and charge this account for it. Clones a voice from one recording with ElevenLabs Instant Voice Cloning, and reads a line in it. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_elevenlabs_voice_design
ChatGPTRun ElevenLabs Voice Design, and charge this account for it. ElevenLabs Voice Design is ElevenLabs' model for creating a brand new voice from a written description. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_elevenlabs_voice_isolator
ChatGPTRun ElevenLabs Voice Isolator, and charge this account for it. ElevenLabs Voice Isolator takes the voice out of a recording and leaves the noise behind. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_gemini_3_8_flash_lite_tts
ChatGPTRun Gemini 3.8 Flash-Lite TTS, and charge this account for it. Gemini 3.8 Flash-Lite TTS is Google's fast, cost-efficient text-to-speech model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_gemini_3_8_flash_tts
ChatGPTRun Gemini 3.8 Flash TTS, and charge this account for it. Gemini 3.8 Flash TTS is Google's text-to-speech model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_minimax_music_2_6
ChatGPTRun MiniMax Music 2.6, and charge this account for it. MiniMax Music 2.6 is MiniMax's text-to-music model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_minimax_music_cover
ChatGPTRun MiniMax Music Cover, and charge this account for it. MiniMax Music Cover is MiniMax's AI cover song model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
audio_scribe
ChatGPTRun ElevenLabs Scribe, and charge this account for it. ElevenLabs Scribe is ElevenLabs' speech-to-text model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
create_character
ChatGPTSave a person, a mascot or a product as a character, so it looks and sounds the same in every later shot. Make its reference pictures first with generate, such as a front view, a side view and a full-body view on a plain background, then save their URLs here. A picture link that is not on Visual Sandbox is copied, so it lasts. Free.
create_design
ChatGPTPut a new design on the person's canvas: a web page, a screen, a card, a poster, or a vector drawing as one <svg>. Free; nothing is generated. Read the 'vsb-design' skill pack first. Name every layer with data-name, because the layers are how the person finds and points at the parts.
create_film_project
ChatGPTStart a multi-shot film that is kept between chats: its brief, script and shot list, and where each shot stands. Opens a canvas for the film; send its sandbox_uuid with every keyframe and clip run, so the whole film lands in one place. The plan is written to the canvas as FILM.md. Free.
cut_and_reframe_video
ChatGPTCut the chosen parts of a long video into one short clip, in the order given, reframe each part to 9:16 or another shape, and burn captions from the source's transcript. For Reels, Shorts and TikToks from a podcast, an interview or a talk. It cuts the moments it is given; it does not pick them. Free. Returns a job id at once; call show_media with it, or get_job when the job is done.
delete_character
ChatGPTDelete a character. Its pictures stay in the library; only the saved character goes. It cannot be undone.
delete_design
ChatGPTDelete a design from the person's canvas, with its node. There is no undo. Delete only a design the person asked you to remove, or one you made in this conversation that they do not want.
download_video
ChatGPTSave the video behind a public page, such as a YouTube, TikTok, Instagram or X link, as a file to edit, cut into clips or caption. Up to 2 hours. A site may refuse the server; then the person uploads the file. Only video the person has the right to use. Free. Returns a job id at once; call show_media with it, or get_job when the job is done.
edit_theme
ChatGPTShow a brand theme in a card where the person changes its colors, fonts, logos, notes and pictures and presses Save. Use it to set up a brand or to let the person fix one. Pass draft with what extract_theme found, so the person checks it before anything is saved. The card saves the theme itself; do not call update_theme for it. Free.
enhance_p_image_upscale
ChatGPTRun Pruna Upscale, and charge this account for it. Pruna Upscale is Pruna AI's image upscaler. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
enhance_recraft_remove_background
ChatGPTRun Remove Background, and charge this account for it. Remove Background is Recraft's AI background remover. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
enhance_upscale
ChatGPTRun Topaz Upscale, and charge this account for it. Upscale is Topaz Labs' AI image enhancer. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
estimate_cost
ChatGPTWhat one call to a model would cost this account, in US dollars, with the supplied inputs. Partial inputs give a partial answer rather than an error. Nothing is charged and no job is created.
export_design
ChatGPTA design as a file. html is the page exactly as the canvas draws it. react is one component in a .jsx file that draws the same. svg works only for a drawing, a design that is one <svg> element. Pass layer to export one frame alone as an html page, with the styles it needs; layer='picked' is the frame the person picked on the canvas.
get_balance
ChatGPTCredit left on this account, in US dollars. Check it before a run that costs real money, and when a generate call is refused for want of credit.
get_character
ChatGPTOne character: its reference pictures, its description and its voice. Pass the pictures to generate as reference images and keep the description in the prompt, so the character stays the same.
get_design
ChatGPTRead a design on the canvas: its layers, its version and its HTML. Pass layer to read one layer's HTML in place of the whole page; that is how to change one part of a big design. layer='picked' reads the layer the person picked on the canvas, and says its path: write back to that path, not to 'picked'. get_selection says which design is picked.
get_film_project
ChatGPTOne film in full: brief, script, the shot list with each shot's keyframe, job and clip, the narration, and the first cut. A shot's clip fills in once its job is done. Free.
get_job
ChatGPTThe state of one generation. While it runs, reports progress and how long it has left. Once it finishes, returns the output URLs and what the run cost. Also returns the finished image itself, or four frames spread across a finished video, so the result can be judged rather than guessed at. A run that waits for review reports awaiting_approval until the person presses Generate, then the run that started.
get_preset
ChatGPTOne saved recipe in full: its model, every input it fixes, and the fields it leaves for the caller to fill in.
get_selection
ChatGPTWhat the person has clicked on the canvas right now. Call it the moment they say 'this one', 'that image' or 'the selected one' without sending a picture, because the selection is what they mean. Answers with an empty list when they have selected nothing.
get_skill
ChatGPTRead one skill pack. Call this before writing your first prompt for a family of model: 'vsb-image-prompting' for any image, 'vsb-nano-banana' for the Nano Banana models, 'vsb-ugc-people' for a person in an advert, 'vsb-video' or 'vsb-p-video-prompting' for video, 'vsb-audio' for sound.
get_theme
ChatGPTOne brand theme in full: colors, fonts, logos, notes and reference pictures. Read it before branded work, and put its colors, look and notes into the prompt; pass a logo or a picture as a reference file. Without a name, returns the default theme. Free.
image_flux_2_klein_9b
ChatGPTRun FLUX.2 Klein 9B, and charge this account for it. FLUX.2 Klein 9B is Black Forest Labs' high-fidelity distilled AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_flux_3_image_klein
ChatGPTRun FLUX 3 Image Klein, and charge this account for it. FLUX 3 Image Klein is the compact model of Black Forest Labs' FLUX 3 family. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_flux_fill_pro
ChatGPTRun FLUX Fill Pro, and charge this account for it. FLUX Fill Pro is Black Forest Labs' AI inpainting and outpainting model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_gpt_image_2
ChatGPTRun GPT Image 2, and charge this account for it. GPT Image 2 is OpenAI's AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_gpt_image_2_5_flare
ChatGPTRun GPT Image 2.5 Flare, and charge this account for it. GPT Image 2.5 Flare is OpenAI's fast image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_gpt_image_2_5_sunburst
ChatGPTRun GPT Image 2.5 Sunburst, and charge this account for it. GPT Image 2.5 Sunburst is OpenAI's precision image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_grok_imagine_image_2
ChatGPTRun Grok Imagine 2.0, and charge this account for it. Grok Imagine 2.0 is xAI's image model for generation and editing. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_krea_2_large
ChatGPTRun Krea 2 Large, and charge this account for it. Krea 2 Large is Krea's foundation image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_krea_2_medium
ChatGPTRun Krea 2 Medium, and charge this account for it. Krea 2 Medium is Krea's mid-size foundation image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_krea_3
ChatGPTRun Krea 3, and charge this account for it. Krea 3 is Krea's instruction-following image editor. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_nano_banana
ChatGPTRun Nano Banana, and charge this account for it. Nano Banana is Google's AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_nano_banana_2
ChatGPTRun Nano Banana 2, and charge this account for it. Nano Banana 2 is Google's AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_nano_banana_2_1
ChatGPTRun Nano Banana 2.1, and charge this account for it. Nano Banana 2.1 is Google's fast, efficient AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_nano_banana_2_lite
ChatGPTRun Nano Banana 2 Lite, and charge this account for it. Nano Banana 2 Lite is Google's AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_nano_banana_pro
ChatGPTRun Nano Banana Pro, and charge this account for it. Nano Banana Pro is Google's AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_p_image_try_on
ChatGPTRun P-Image Try-On, and charge this account for it. P-Image Try-On is Pruna AI's purpose-built virtual try-on model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_recraft_v4_1
ChatGPTRun Recraft V4.1, and charge this account for it. Recraft V4.1 is Recraft's design-minded AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_recraft_v4_1_flash
ChatGPTRun Recraft V4.1 Flash, and charge this account for it. Recraft V4.1 Flash is Recraft's image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_seedream_5_lite
ChatGPTRun Seedream 5 Lite, and charge this account for it. Seedream 5 Lite is ByteDance's fast AI image model for generation and editing. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_seedream_5_pro
ChatGPTRun Seedream 5 Pro, and charge this account for it. Seedream 5 Pro is ByteDance's AI image model for generation and editing. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
image_z_image_turbo
ChatGPTRun Z-Image Turbo, and charge this account for it. Z-Image Turbo is Tongyi-MAI's sub-second AI image model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
import_media
ChatGPTShow a card where the person picks a picture, a video or a sound from their device. The card uploads it and sends you its lasting URL in a message. Wait for that message, then pass the URL to a model as a reference. Use it when the work needs the person's own file and they have not attached one. Free.
join_clips
ChatGPTJoin video clips and pictures into one video, in order, with a cut or a fade between them. Every clip is fitted to one shape, so shots from different models line up. Use it to build an ad, a reel or a short film from the shots you generated. Free. Returns a job id at once; call show_media with it, or get_job when the job is done.
list_characters
ChatGPTThe characters saved on this account, with their reference pictures and voices. Call it when the person names someone they made before.
list_film_projects
ChatGPTThe film projects on this account, most recently changed first. Free.
list_jobs
ChatGPTRecent generations on this account, newest first. Use it to find a job id from an earlier session, or to see what is still running.
list_presets
ChatGPTSaved recipes: a model plus the inputs and reference images that reproduce a particular look. Run one by reading it with get_preset and passing its inputs to generate.
list_sandbox_items
ChatGPTEverything on the canvas the person is watching: the pictures, videos and uploads already sitting there, newest first, each with the URL to pass back to generate as a reference. Call it when the person points at something they can see but have not sent you, or to check what the work has already produced.
list_skills
ChatGPTThe Visual Sandbox skill packs: written guidance on how to prompt each family of model well. Read the one that fits before writing a prompt. Prompt craft changes the result far more than the choice of model does.
list_themes
ChatGPTThe brand themes on this account: each one's name, whether it is the default, and its thumbnail. Free.
list_voices
ChatGPTThe voices this account made with ElevenLabs Voice Design or Voice Clone. To speak in one, call generate on audio/elevenlabs-tts with its name or id as my_voice.
mesh_hunyuan_3d_3_1
ChatGPTRun Hunyuan 3D 3.1, and charge this account for it. Hunyuan 3D 3.1 is Tencent's AI 3D model generator. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
mesh_meshy_7_1
ChatGPTRun Meshy 7.1, and charge this account for it. Meshy 7.1 is Meshy's AI 3D model generator. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
mesh_rodin_gen_2_5
ChatGPTRun Rodin Gen-2.5, and charge this account for it. Rodin Gen-2.5 is Hyper3D's AI 3D model generator. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
mesh_sam_3_3d_objects
ChatGPTRun SAM 3D Objects, and charge this account for it. SAM 3D Objects is Meta's AI 3D model generator, built for real photos. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
mesh_trellis_2
ChatGPTRun TRELLIS 2, and charge this account for it. TRELLIS 2 is Microsoft's open source AI 3D model generator. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
mesh_tripo_p2
ChatGPTRun Tripo P2, and charge this account for it. Tripo P2 is Tripo's AI 3D model generator, built for games and real-time pipelines. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
move_layer
ChatGPTRearrange a design: move one layer just before or just after another layer, or last inside one. Only its place in the HTML changes; nothing else is rewritten. Pass exactly one of before, after or into.
read_sandbox_docs
ChatGPTThe markdown files the person keeps on this canvas — BRAND.md, CHARACTER.md, a shot list — with every file's contents. They hold what the work has to be: the palette, the voice, the face that has to stay the same across twenty shots. Read them before you write a prompt, so the twentieth picture matches the first. Answers with an empty list when the person keeps no notes.
render_html_image
ChatGPTRender an HTML layout to a PNG with exact text: a social card, a quote card, a thumbnail, a chart, a mockup, or a logo or watermark on a transparent background. Pick a moment of an animation with time. Free. Returns a job id at once; call show_media with it, or get_job when the job is done.
render_html_video
ChatGPTRender an HTML composition to an MP4: motion graphics, kinetic typography, an animated explainer, a title card, lower thirds or a logo sting. You write the HTML; every frame is drawn at an exact time, so the same HTML always gives the same video. Up to 60 seconds. Free. Returns a job id at once; call show_media with it, or get_job when the job is done.
search_assets
ChatGPTFind finished results and uploads in the person's library: a logo, a product photo, an approved keyframe or an earlier take. Use it when the request names saved media or needs the same product, person or brand again. Skip it for a new picture that needs nothing saved. Free.
show_media
ChatGPTShow generations to the person as a card in the chat: the picture, the video or the sound with a player, and the percent of a job that still runs. Call it once, right after a run or an edit, with every job id of the request. The card follows each job until it ends, so there is no need to poll get_job for the person to see the result. Call get_job only to look at a result yourself. A run that waits for review shows its request and price with a Generate button; the run starts only when the person presses it.
submit_feedback
ChatGPTSend product feedback to the Visual Sandbox team: a bug, a result that missed what was asked, a missing feature or model, something confusing, or praise. Send what the person says, or a concrete problem you saw in this task, once for each issue. Tell the person in one line that you sent it, and do not send it when they ask you not to. It is not a refund request and starts no new run. Free.
transcribe_link
ChatGPTTranscribe the video behind a public page, such as a YouTube, TikTok, Instagram or X link, with a start and an end time for every word. A page's own captions are read for free. With use_captions false, ElevenLabs Scribe transcribes the sound at its price for the real length. Free with the page's captions. With use_captions false, Scribe's price for the real length; Visual Sandbox holds the price of 2 hours while it runs. Returns a job id at once; call show_media with it, or get_job when the job is done.
trim_video
ChatGPTCut the part you want out of a video, reframe it to another shape such as 9:16 for TikTok and Reels, and speed it up or slow it down. Any of the three, in one step. Free. Returns a job id at once; call show_media with it, or get_job when the job is done.
update_asset
ChatGPTGive an asset a name, tags or a description, so it can be found again with search_assets. Names are unique on the account, such as 'acme-logo-color' or 'launch-shot-03-final'. tags replaces every tag the asset has, so list the ones to keep. A file uploaded with upload_media joins the library when its URL is named here. Free.
update_character
ChatGPTChange a character: rename it, rewrite its description, replace its reference pictures or give it a voice. Only the fields given change.
update_design
ChatGPTChange a design on the canvas. With layer, html replaces that one layer (the whole element, data-name and all); without it, html replaces the whole page. Send the version you read: when the person changed the design since, the write is refused and you read it again.
update_film_project
ChatGPTChange a film. Each field given replaces that field. shots replaces the whole shot list; patch_shots changes shots by id and keeps the rest (a new id adds a shot at the end, remove: true drops one). Record a shot's job_id when its run starts; its clip fills in by itself. Move the stage on only when the person approved that step. Free.
update_theme
ChatGPTCreate a brand theme or change one. Each field given replaces that field; a field left out keeps its value. The first theme becomes the default. A logo or a picture on another site is copied, so it lasts. Save only values the person gave or confirmed: show a guess first. Free.
upload_media
ChatGPTPut a file on the Visual Sandbox CDN and return a lasting URL to pass to generate as a reference: a picture, a video, a sound or a 3D model, up to 100MB. Give exactly one of: file, for a file the person attached in the chat; data_uri, for a local file; url, to keep a copy of a public file. A public https URL can also go to generate directly.
vector_quiver_arrow_2_animate
ChatGPTRun Quiver Arrow 2 to animate an SVG, and charge this account for it. Quiver Arrow 2 is Quiver AI's AI vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_quiver_arrow_2_edit
ChatGPTRun Quiver Arrow 2 to change an SVG with a prompt, and charge this account for it. Quiver Arrow 2 is Quiver AI's AI vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_quiver_arrow_2_generate
ChatGPTRun Quiver Arrow 2 to make an SVG from a prompt, and charge this account for it. Quiver Arrow 2 is Quiver AI's AI vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_quiver_arrow_2_telos_animate
ChatGPTRun Quiver Arrow 2 Telos to animate an SVG, and charge this account for it. Quiver Arrow 2 Telos is Quiver AI's deep-thinking AI vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_quiver_arrow_2_telos_edit
ChatGPTRun Quiver Arrow 2 Telos to change an SVG with a prompt, and charge this account for it. Quiver Arrow 2 Telos is Quiver AI's deep-thinking AI vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_quiver_arrow_2_telos_generate
ChatGPTRun Quiver Arrow 2 Telos to make an SVG from a prompt, and charge this account for it. Quiver Arrow 2 Telos is Quiver AI's deep-thinking AI vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_quiver_arrow_2_telos_vectorize
ChatGPTRun Quiver Arrow 2 Telos to turn a picture into an SVG, and charge this account for it. Quiver Arrow 2 Telos is Quiver AI's deep-thinking AI vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_quiver_arrow_2_vectorize
ChatGPTRun Quiver Arrow 2 to turn a picture into an SVG, and charge this account for it. Quiver Arrow 2 is Quiver AI's AI vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_recraft_v4_1_svg
ChatGPTRun Recraft V4.1 SVG, and charge this account for it. Recraft V4.1 SVG is Recraft's text-to-vector model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
vector_recraft_vectorize
ChatGPTRun Recraft Vectorize, and charge this account for it. Recraft Vectorize is Recraft's purpose-built raster-to-SVG converter. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_flux_3
ChatGPTRun FLUX 3, and charge this account for it. FLUX 3 is Black Forest Labs' multimodal video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_flux_video_upscale
ChatGPTRun FLUX Video Upscale, and charge this account for it. FLUX Video Upscale is Black Forest Labs' AI video upscaler. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_grok_imagine_video
ChatGPTRun Grok Imagine Video, and charge this account for it. Grok Imagine Video is xAI's video model with sound built in. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_grok_imagine_video_1_5
ChatGPTRun Grok Imagine Video 1.5, and charge this account for it. Grok Imagine Video 1.5 is xAI's video model, and it animates a picture. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_kling_avatar_v2
ChatGPTRun Kling Avatar 2.0, and charge this account for it. Kling Avatar 2.0 is Kuaishou's talking-head avatar model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_kling_v3_motion_control
ChatGPTRun Kling 3.0 Motion Control, and charge this account for it. Kling 3.0 Motion Control transfers motion from a reference video onto a still character image. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_kling_v3_omni_video
ChatGPTRun Kling 3.0 Omni, and charge this account for it. Kling 3.0 Omni is Kuaishou's unified multimodal video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_lipsync_2_pro
ChatGPTRun Lipsync 2 Pro, and charge this account for it. Lipsync 2 Pro is Sync's studio-grade lip sync model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_ltx_2_5_fast
ChatGPTRun LTX 2.5 Fast, and charge this account for it. LTX 2.5 Fast is Lightricks' speed tier for open-weights AI video. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_minimax_h3
ChatGPTRun MiniMax H3, and charge this account for it. MiniMax H3 is MiniMax's multimodal video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_p_video
ChatGPTRun P-Video, and charge this account for it. P-Video is Pruna AI's purpose-built fast video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_p_video_2
ChatGPTRun P-Video 2, and charge this account for it. P-Video 2 is Pruna AI's quality-focused video model, the successor to P-Video. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_p_video_2_pro
ChatGPTRun P-Video 2 Pro, and charge this account for it. P-Video 2 Pro is Pruna AI's video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_p_video_animate
ChatGPTRun P-Video Animate, and charge this account for it. P-Video Animate is Pruna AI's motion transfer model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_p_video_avatar
ChatGPTRun P-Video Avatar, and charge this account for it. P-Video Avatar is Pruna AI's purpose-built talking-head video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_p_video_edit
ChatGPTRun P-Video Edit, and charge this account for it. P-Video Edit is Pruna AI's video editing model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_p_video_replace
ChatGPTRun P-Video Replace, and charge this account for it. P-Video Replace is Pruna AI's person swap model for video. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_seedance_2
ChatGPTRun Seedance 2.0, and charge this account for it. Seedance 2.0 is ByteDance's AI video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_seedance_2_0_mini
ChatGPTRun Seedance 2.0 Mini, and charge this account for it. Seedance 2.0 Mini is the low-cost tier of ByteDance's Seedance 2.0 video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_seedance_2_5
ChatGPTRun Seedance 2.5, and charge this account for it. Seedance 2.5 is ByteDance's AI video model and the successor to Seedance 2.0. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_seedance_2_fast
ChatGPTRun Seedance 2.0 Fast, and charge this account for it. Seedance 2.0 Fast is the speed tier of ByteDance's Seedance 2.0 video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_starlight_fast_2
ChatGPTRun Topaz Starlight Fast 2, and charge this account for it. Topaz Starlight Fast 2 is Topaz Labs' generative AI video upscaler, built for speed. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_starlight_precise_2_6
ChatGPTRun Topaz Starlight Precise 2.6, and charge this account for it. Topaz Starlight Precise 2.6 is Topaz Labs' generative AI video upscaler, designed for detail. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_veo_3_1
ChatGPTRun Veo 3.1, and charge this account for it. Veo 3.1 is Google DeepMind's AI video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_veo_3_1_fast
ChatGPTRun Veo 3.1 Fast, and charge this account for it. Veo 3.1 Fast is the speed tier of Google DeepMind's Veo 3.1 video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_video_upscale
ChatGPTRun Topaz Video Upscale, and charge this account for it. Topaz Video Upscale is Topaz Labs' AI video upscaler. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
video_wan_3
ChatGPTRun Wan 3.0, and charge this account for it. Wan 3.0 is Alibaba's AI video model. A run of $0.10 or more can wait for the person to press Generate on the card, and nothing is charged before. Returns a job id at once, not the media; call show_media with it.
voice_studio
ChatGPTShow the takes of a voice in a card, so the person hears them side by side, picks one and saves it to a character. First run audio_elevenlabs_voice_design with auditions 3 and a sample line of 100 to 400 characters, so the run costs a few cents and starts at once; then call this with its job id. The card plays the takes when they are ready. On an account that may clone a voice, the card also records a sample: it then sends you the recording's URL with the person's consent, and you run audio_elevenlabs_voice_clone. The card saves the pick itself; do not call update_character for it. Free.
write_sandbox_doc
ChatGPTWrite one of the person's markdown files on this canvas, by name. Use it when they settle on something worth keeping — a palette, a character, a rule the work has to follow — so the next run and the next session still have it. It replaces the whole file, so read it first and send back the version you want kept, or you will drop what is already there.
