Every tool the server gives your assistant
The Makemancer MCP server lists 77 tools in Claude, Claude Code and Codex, and a ChatGPT set without purchase tools. You never call them by hand: ask in plain words and the assistant picks the tool. This page is generated from the same tool list the server sends, with every parameter.
Generate
generate_image
Generate imageChanges dataAsk for it like thisMake a product photo of my perfume on travertine, 4:5.
Parameters (12)
Name Type What it is promptrequiredstring What to generate, or the edit instructions when image_url is given. modelstring Image model ID. Default image/nano-banana-2. image_urlstring Source image to edit (https URL). image_urlsstring[] Reference images (https URLs). character_idsstring[] Saved Characters for this shot, in order; a one-item array for one Character. Use list_characters for names and ids, then use the exact saved names in prompt. The server adds each identity description and reference image after existing references and states their Image N numbers; do not guess those numbers. All references count toward the model limit; unsupported requests fail before charging. aspect_ratiostring Output shape such as 16:9; allowed values per model in get_model_params. size"1:1" | "4:5" | "5:4" | "9:16" | "16:9" Legacy alias of aspect_ratio. num_imagesinteger Number of images to generate (1-4). Each image is its own job, charged one image's price and refunded on its own if it fails; they share one job_id and the card shows them side by side. agentic_promptingboolean Optional prompt rewrite for the selected model. Omit it (default) or set false to preserve the user's wording; set true only when the user wants a model-specific rewrite. Adds a few seconds and costs nothing extra; the result returns the rewritten prompt (and original_prompt), or a warning when the original was kept. Selected Character identity context is appended unchanged either way. extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. generate_video
Generate videoChanges dataAsk for it like thisAnimate this photo into a 5-second vertical clip.
Parameters (15)
Name Type What it is promptrequiredstring Subject, action, camera, setting, timing and sound of the clip. modelstring Video model ID. Default video/gemini-omni-flash. image_urlstring Literal first frame (https URL). last_frame_urlstring Last frame, for models that support it. image_urlsstring[] Reference images that guide but do not become frames. character_idstring A saved Character. Its description is added to the prompt and its image is appended to image_urls as a reference, never as a start frame, so it cannot be combined with image_url or last_frame_url. Choose a model that supports image references. It does not select the Character's voice for native video audio. Use list_characters to find ids. video_urlsstring[] Reference or source videos. audio_urlsstring[] Reference audio, for models that support it. durationnumber | string Length in seconds, e.g. 5 or "5s"; limits per model. aspect_ratiostring e.g. 16:9, 9:16, 1:1. resolutionstring e.g. 480p, 720p, 1080p; allowed values per model. agentic_promptingboolean Optional prompt rewrite for the selected model. Omit it by default to preserve the user's wording while keeping the provider's own prompt expansion where a model has one. Set true only when the user wants a model-specific rewrite; the result returns the rewritten prompt (and original_prompt), or a warning when the original was kept. Set false for strict wording; it also turns off the provider's prompt expansion. Adds a few seconds and costs nothing extra. Selected Character identity context is appended unchanged either way. extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. generate_speech
Generate speechChanges dataAsk for it like thisRead this script as a calm narrator.
Parameters (9)
Name Type What it is textstring Text to speak. Some models take dialogue turns in extras instead. modelstring Speech model ID. Default speech/elevenlabs-v4. voice_idstring Stock voice ID from get_model_params or search_voices. For a cloned or designed voice pass character_id instead. character_idstring Character with a saved voice for this model (see list_characters). Do not combine with voice_id. In extras.dialogue, a turn may also take character_id. voice_option_idstring A saved voice on that Character. Omit to use the preferred voice for the selected model. apply_saved_pronunciationsboolean Apply this account's saved pronunciation text replacements for the selected model family. Defaults to true. Set false to bypass saved entries for this request; explicit pronunciation markup in text is preserved. extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. generate_music
Generate musicChanges dataAsk for it like thisWrite a 20-second acoustic jingle for my brand.
Parameters (14)
Name Type What it is promptstring Genre, mood, instruments, tempo, structure and vocal role. modelstring Music model ID. Default music/standard. lyricsstring Custom lyrics; section tags such as [verse] on their own lines. stylestring Style tags for custom mode. titlestring Track title for custom mode. instrumentalboolean No vocals. Default false. force_instrumentalboolean Alias of instrumental. music_length_msinteger Approximate length in milliseconds (converted to duration in seconds). audio_urlstring Uploaded source song for music/standard-cover. audio_idstring A generated track to extend, cover or edit: audio_id from the tracks of a completed music result. task_idstring task_id from the same tracks entry (the job_id of that result). extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. generate_sound_effect
Generate sound effectChanges dataAsk for it like thisMake 4 seconds of rain on a window.
Parameters (8)
Name Type What it is promptstring The sound: source, action, space, texture and ending. modelstring Sound-effect model ID. Default sfx/elevenlabs-sound-v2. duration_secondsnumber Length in seconds; omit to let the model choose. loopboolean Make a seamless loop. prompt_influencenumber How closely to follow the prompt, 0 to 1. extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project.
Edit
remove_background
Remove backgroundChanges dataAsk for it like thisCut the product out of this photo.
Parameters (7)
Name Type What it is media_urlrequiredstring Image or video to process (https URL). typerequired"image" | "video" Kind of media at media_url. modelstring Background-removal model ID. output_format"webm" | "mp4" | "prores" Video only. extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. upscale_image
Upscale imageChanges dataAsk for it like thisMake this image sharper and bigger.
Parameters (5)
Name Type What it is image_urlrequiredstring Image to upscale (https URL). modelstring Image upscaler model ID. extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. upscale_video
Upscale videoChanges dataAsk for it like thisUpscale this clip to 1080p.
Parameters (5)
Name Type What it is video_urlrequiredstring Video to upscale (https URL). modelstring Video upscaler model ID. extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project.
Plan and price
estimate_generation
Estimate generation costRead-onlyAsk for it like thisHow much would a 10-second Kling clip cost?
Parameters (3)
Name Type What it is operation"image" | "video" | "speech" | "audio" | "upscale" | "video_upscale" | "html_video" | "frame_interpolation" | "lip_sync" | "swap_video_characters" | "3d_model" | "video_edit" | "media_import" | "transcribe" Which generation the estimate is for: image, video, speech, audio (music, or a sfx/ model), upscale, video_upscale, transcribe, media_import or html_video (params.method render, or generate / edit for generate_html_video). Optional when model is given. For social URLs, media_import estimates upload_asset with params url, type audio/video, quality and max_duration, and transcribe estimates social caption retrieval including its automatic audio fallback; both read the source length from metadata without downloading and return the full prepaid reservation. frame_interpolation estimates increase_frame_rate (params video_url plus extras target_fps, slowdown_factor) and lip_sync estimates lip_sync_video (params video_url, plus audio_url and model when given). swap_video_characters quotes its price with its free previews, so call it directly. modelstring Model ID. Omit to use the default model of the operation. paramsobject The arguments the generation tool would receive (top-level fields and extras). list_models
List modelsRead-onlyAsk for it like thisWhich video models can you use?
Parameters (2)
Name Type What it is category"image" | "video" | "speech" | "audio" | "music" | "sfx" | "upscale" | "background_removal" | "frame_interpolation" Filter by category. Omit to list all. querystring Free-text filter on ID, name, description and capabilities. get_model_params
Get model parametersRead-onlyAsk for it like thisWhat settings does Seedance 2.5 take?
Parameters (2)
Name Type What it is modelrequiredstring Model ID, e.g. image/nano-banana-2 or video/seedance-2.5. include_voice_catalogboolean Accepted for compatibility; voice IDs are part of the schema.
Jobs and results
check_job
Check jobRead-onlyAsk for it like thisIs my video ready yet?
Parameters (2)
Name Type What it is job_idrequiredstring job_id returned by a generation tool (or its first eight characters), a cloud_project_id or a swap session_id. words_offsetinteger transcribe: next page of words; pass nextWordsOffset. show_media
Show mediaRead-onlyAsk for it like thisShow me the last image again.
Parameters (1)
Name Type What it is itemsrequiredobject[] Media to show; each item has one of asset_id, job_id or url.
Your library and uploads
search_assets
Search assetsRead-onlyAsk for it like thisFind the olive oil photos I made last week.
Parameters (11)
Name Type What it is type"image" | "video" | "audio" | "document" | "zip" | "project" | "3d_model" | "font" Asset type. querystring Matches the asset name (substring), one tag, or the prompt of the generation that made it (substring). tagsstring[] Assets with any of these tags. namestring Exact asset name. source"generated" | "uploaded" | "imported" How the asset entered the library. created_fromstring Created at or after this time (ISO 8601). created_tostring Created before this time (ISO 8601). limitinteger Results per page. offsetinteger Results to skip. favorites_onlyboolean Not supported; only false is accepted. budget_projectstring Only generated work in this budget project (its slug, id or name), or "none" for generated work in no budget project. Each result names its budget project when it has one. update_asset
Update assetDeletesAsk for it like thisRename that clip to “launch teaser”.
Parameters (5)
Name Type What it is idrequiredstring Asset ID. namestring | null New name; null clears it. clear_nameboolean Remove the name. tagsstring[] Replaces all tags. descriptionstring Accepted for compatibility; not stored. delete_asset
Delete assetDeletesAsk for it like thisDelete the draft video.
Parameters (1)
Name Type What it is idrequiredstring Asset ID. upload_asset
Import media from URLChanges dataAsk for it like thisSave this image link to my library.
Parameters (8)
Name Type What it is urlrequiredstring Public direct file URL or supported social post URL to import and store permanently. Social imports return a job ID for check_job. namestring Asset name; defaults to the file name. tagsstring[] Tags for organizing the asset. type"image" | "video" | "audio" | "3d_model" | "zip" Expected type of a direct file; the import fails if the file is another type. Required for a ZIP (zip) or 3D model (3d_model) served with a generic type such as application/octet-stream. For social links, omit for video or use audio for audio-only downloads. quality"480p" | "720p" | "1080p" Social video resolution: 480p (default, 1.2 credits per started minute), 720p or 1080p (4.8). Total credits round up once per import. Ignored for type audio. max_durationinteger For social links, reject longer sources before downloading. Default 7200 seconds. This rejects long sources; it does not trim them. content_typestring Accepted for compatibility; the type is read from the file. Omit for social links. descriptionstring Accepted for compatibility; not stored. get_upload_url
Get upload URLChanges dataAsk for it like thisUpload this file from my computer.
Parameters (8)
Name Type What it is content_typerequiredstring MIME type, e.g. image/png, video/mp4, audio/mpeg, font/woff2, application/zip. With a generic type (application/octet-stream) the filename extension decides. typerequired"image" | "video" | "audio" | "3d_model" | "font" | "zip" | "project" Asset type; must match content_type. zip is an ordinary ZIP asset; project is a render_html_video project ZIP, kept private and reused by asset ID. bytesinteger Exact file size in bytes, when known; it is then part of the signed upload. confirm_upload checks the size limit either way. namestring Asset name. filenamestring Original file name: picks the type when the browser reports a generic content_type (e.g. application/octet-stream for woff2), and is the name when name is omitted. tagsstring[] Tags for organizing the asset. descriptionstring Accepted for compatibility; not stored. purpose"media" | "voice_clone" Use voice_clone for a private voice recording; returns an asset ID for clone_voice, never a public audio URL. confirm_upload
Confirm uploadChanges dataAsk for it like this(called after an upload finishes)
Parameters (1)
Name Type What it is asset_idrequiredstring asset_id returned by get_upload_url. load_image
Load imageRead-onlyAsk for it like thisLook at the image I just made.
Parameters (1)
Name Type What it is urlrequiredstring Public image URL. import_chatgpt_media
Import ChatGPT attachmentChanges dataChatGPT onlyAsk for it like thisUse the photo I attached.
Parameters (5)
Name Type What it is media_filerequiredobject The ChatGPT file to import. namestring Asset name; defaults to the file name. tagsstring[] Tags for organizing the asset. type"image" | "video" | "audio" Expected type; the import fails if the file is another type. purpose"media" | "voice_clone" voice_clone keeps an audio recording private for clone_voice and returns its asset ID, never a public URL.
Account
manage_account
View accountRead-onlyAsk for it like thisWhat did I spend this month?
Parameters (1)
Name Type What it is section"overview" | "activity" | "settings" Account section to show: overview, activity or settings. get_credits_balance
Get credit balanceRead-onlyAsk for it like thisHow many credits do I have?
Parameters (3)
Name Type What it is estimate_type"image" | "video" | "speech" | "audio" | "upscale" | "video_upscale" | "html_video" | "frame_interpolation" | "lip_sync" | "swap_video_characters" | "3d_model" | "video_edit" | "media_import" | "transcribe" Operation to estimate, as in estimate_generation (image, video, speech, html_video…). video_edit needs estimate_model; 3d_model is not available on this server and says so. estimate_modelstring Model ID to estimate. estimate_paramsobject Arguments of the generation to estimate. get_credits_link
Get credits linkRead-onlyNot in ChatGPTAsk for it like thisI need more credits.
Parameters (1)
Name Type What it is product"1000" | "5000" | "11000" | "27500" | "62500" | "credits_1000" | "credits_5250" | "credits_11000" | "credits_28750" Pack to highlight; omit to let the user choose. submit_feedback
Submit feedbackChanges dataAsk for it like thisTell the team the last result had the wrong aspect ratio.
Parameters (6)
Name Type What it is messagerequiredstring What worked, broke, was missing or was confusing. categoryrequired"bug" | "generation_quality" | "missing_feature" | "confusing" | "praise" | "other" Kind of feedback. sourcerequired"agent" | "user" agent: observed by the assistant; user: the user's own words. attemptedTaskstring What the user was trying to do. toolNamestring Tool involved, if any. relatedIdsstring[] Related job or asset IDs.
Other
add_audio_to_video
Add Audio To VideoChanges dataParameters (8)
Name Type What it is video_urlrequiredstring Video to put the audio on. audio_urlrequiredstring Audio to add, such as narration or music. start_offsetnumber Seconds after the video starts when the audio begins. Default: 0. audio_mode"replace" | "mix" Default: replace. original_volumenumber With mix: gain of the video's own audio, 0 to 1. Default: 1. added_volumenumber With mix: gain of audio_url, 0 to 1. Default: 1. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. add_speech_pronunciations
Save speech pronunciationsDeletesParameters (1)
Name Type What it is entriesrequiredobject[] Up to 100 pronunciations to save. add_subtitles
Add SubtitlesChanges dataParameters (17)
Name Type What it is video_urlrequiredstring Video to caption. languagestring ISO-639 code, sent only when known. Speech language is detected automatically; this picks the font for Chinese, Japanese or Korean. font_namestring Exact Google Fonts family name, only when the user names a font; never choose one. Checked for the language's script before charging. Words in a script the font cannot draw use the matching Noto Sans family. font_sizeinteger Preferred size in pixels; kept within a safe range for the frame. font_weight"normal" | "bold" | "black" Default: bold. A font without the weight uses its nearest one. font_color"white" | "black" | "red" | "green" | "blue" | "yellow" | "orange" | "purple" | "pink" | "brown" | "gray" | "cyan" | "magenta" Default: white. highlight_color"white" | "black" | "red" | "green" | "blue" | "yellow" | "orange" | "purple" | "pink" | "brown" | "gray" | "cyan" | "magenta" Color of the word being spoken, the karaoke cue. Default: white. stroke_color"white" | "black" | "red" | "green" | "blue" | "yellow" | "orange" | "purple" | "pink" | "brown" | "gray" | "cyan" | "magenta" Default: black. stroke_widthinteger Outline in pixels, 0 for none. Default: 3. background_color"white" | "black" | "red" | "green" | "blue" | "yellow" | "orange" | "purple" | "pink" | "brown" | "gray" | "cyan" | "magenta" | "none" | "transparent" Box behind the text. Default: none. background_opacitynumber Default: 0.7. position"top" | "center" | "bottom" Default: bottom. y_offsetinteger Top and bottom: inward margin, 20 to 200. Center: pixels below the middle, -200 to 200. words_per_subtitleinteger Preferred most words per caption; limited by the frame shape. enable_animationboolean Bounce each caption in. Default: false. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. approve_generation
Approve generationChanges dataParameters (4)
Name Type What it is request_idrequiredstring Review request ID. expected_versionrequiredinteger The version shown to the user. promptstring The prompt as edited in the card. settingsobject Resolution, duration, aspect ratio and audio as chosen. assemble_film
Assemble FilmChanges dataParameters (5)
Name Type What it is idrequiredstring Existing film project id returned by create_film_project or list_film_projects; never use a placeholder. mode"connect" | "cut_end" connect preserves every clip and warns about target overage; cut_end trims the fully assembled output at the target duration. video_fit"auto" | "crop" | "pad" | "strict" Fit clips to the first clip's canvas. crop (default) and auto center-crop mismatches; pad preserves the full frame with black bars; strict (only when the user asks for it) refuses clips with a different aspect ratio. Original shot URLs are preserved; completed crops are reused on later assemblies. with_narrationboolean Replace the stitched video's audio with the project's audioUrl narration track, if present. This discards every shot's own audio, and the cut ends with the shorter of the video and narration. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. assign_to_budget_project
Assign To Budget ProjectChanges dataParameters (3)
Name Type What it is budget_projectrequiredstring | null Target project's slug, id, or name; null removes the generations from their project. job_idsstring[] Generation job ids to move. asset_idsstring[] Asset ids; each moves its generation. clone_voice
Clone voiceDeletesParameters (7)
Name Type What it is audio_asset_idstring Private voice recording ID from an upload with purpose voice_clone. Public media assets cannot be used. audio_urlstring A publicly downloadable audio link. The original bytes are copied into this account's private voice storage before cloning. Private-network targets and non-audio files are rejected. Prefer audio_asset_id when the recording is already privately uploaded. character_idstring Existing Character to receive the cloned voice. Omit to create a reusable voice-only Character automatically. character_namestring Name for an automatically created voice-only Character. Used only when character_id is omitted; defaults to "My Voice". consentboolean Pass only with audio_asset_id or audio_url, as true after the user agreed to the consent statement: the recording is their own voice or the speaker explicitly permitted cloning and speech generation. Omit it when calling without a recording. languagestring Source recording language code, such as en, es, or he. provider"elevenlabs" | "cartesia" | "fish" Voice provider. Here only fish is connected and it is the default; providers are never switched silently, so cartesia or elevenlabs is refused. create_budget_project
Create Budget ProjectChanges dataParameters (3)
Name Type What it is namerequiredstring Display name, e.g. Papercraft Tarot. Unique in the workspace. slugstring Optional kebab-case slug. Generated from the name when omitted. notesstring Optional notes, such as the client or a purchase order. create_character
Create CharacterChanges dataParameters (9)
Name Type What it is titlerequiredstring Character name. descriptionstring Appearance, personality, role: added to prompts at generation time. image_urlstring Reference portrait or character sheet URL: the visual anchor for generate_image and generate_video. profile_image_urlstring Small browsing portrait, stored separately from image_url. If omitted when adding or changing a sheet, a profile portrait is generated asynchronously with normal image credits. For a portrait-only Character, pass the portrait URL as both image_url and profile_image_url. voice_modelstring Speech model of an existing public voice, such as speech/elevenlabs-v4. voice_idstring Exact public voice ID for voice_model, from get_model_params. This does not create a new voice. voice_namestring Optional display name for the saved public voice. voice_preview_idstring preview_id of a design_voice audition to save as a voice of this Character. It becomes a private Fish voice that speaks with speech/fish-drama-3. voice_languagestring | null Optional voice preview language code, such as en, he or es. Null clears it. With a saved voice, setting or changing language generates one short localized preview asynchronously using normal speech credits. Omit to keep the current value; no language means no automatic preview. create_film_project
Create Film ProjectChanges dataParameters (6)
Name Type What it is namerequiredstring Short project title briefstring The user's creative brief / what the film is about character_idsstring[] Character ids for the cast (from list_characters) theme_idstring Brand theme id target_duration_snumber Target total film length in seconds, up to 60 minutes; shot duration is model-dependent. Fractional values are rounded to the nearest second. budget_projectstring Tie this film to a budget project (slug, id or name). Its shot clips and assembly count toward it. create_template
Create TemplateChanges dataParameters (13)
Name Type What it is namerequiredstring Template name. descriptionstring What this template is for. kind"html" | "generative" Default: html. parametersobject[] The {{token}} inputs: name, type, optional default, description and required. htmlstring html kind: markup with {{token}} placeholders. widthinteger html kind: default width. Required unless sizes is given. heightinteger html kind: default height. Required unless sizes is given. sizesobject[] html kind: named output sizes; one may set isDefault. media_type"image" | "video" generative kind: output type. recommended_modelstring generative kind: model ID. promptstring generative kind: prompt with {{token}} placeholders. model_paramsobject generative kind: default model parameters. reference_asset_idsstring[] generative kind: Library assets passed as references. cut_and_reframe_video
Cut And Reframe VideoChanges dataParameters (12)
Name Type What it is video_urlrequiredstring A video in your Library; import the file first. segmentsrequiredobject[] Half-open source ranges in seconds, in playback order. widthinteger Output width, even, 128-1920. Default: 1080. heightinteger Output height, even, 128-1920. Default: 1920. fps"24" | "25" | "30" | "50" | "60" | "source" Default: source. audio_fade_msnumber Fade at each discontinuous audio cut, milliseconds. Default: 30. Above 100 is treated as 100. normalize_audioboolean Level the whole output. Default: false. captionsobject Optional burned captions. namestring Name for the result in the Library. tagsstring[] Tags for the result. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. delete_character
Delete CharacterDeletesParameters (2)
Name Type What it is idrequiredstring ID of the Character to delete. confirmrequiredboolean Set true only after the user asked to delete this Character. Deletion cannot be undone. delete_film_project
Delete Film ProjectDeletesParameters (4)
Name Type What it is idrequiredstring Film project ID to permanently delete. confirmboolean True only after the user explicitly confirms deletion of this film. delete_mediaboolean Also delete currently linked owned media, only if explicitly requested. Defaults to keeping media. expected_updated_atstring The project's updatedAt from the preview. Rejects a stale deletion confirmation. delete_speech_pronunciations
Delete speech pronunciationsDeletesParameters (1)
Name Type What it is idsrequiredstring[] Entry IDs from list_speech_pronunciations. design_voice
Design voiceChanges dataParameters (13)
Name Type What it is promptrequiredstring Voice description, 20 to 2,000 characters: age, accent, timbre, pacing, tone and speaking style. textstring Audition script spoken by all three candidates: 1 to 150 characters for Fish. Omit when no script exists yet. provider"elevenlabs" | "google" | "fish" Voice Design provider. Only fish is connected here and it is the default; elevenlabs and google are refused rather than switched silently. extrasobject Fish controls: seed, language_code (BCP-47), guidance_scale (default 2), instruct_guidance_scale (default 0), speed (0 < speed <= 3, default 1), num_step (1 to 128, default 32). Unsupported controls are ignored. They are also accepted at the top level; extras take precedence. language_codestring BCP-47 audition language, such as en-GB or he-IL (same as extras.language_code). guidance_scalenumber Same as extras.guidance_scale. instruct_guidance_scalenumber Same as extras.instruct_guidance_scale. speednumber Same as extras.speed. num_stepinteger Same as extras.num_step. seedinteger Same as extras.seed. character_idstring Existing Character the voice is meant for. It is remembered as the place to save; this call saves nothing. character_namestring Name of a new Character the voice is meant for, when character_id is omitted. It labels the options and is remembered for saving. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. extract_frames
Extract FramesChanges dataParameters (12)
Name Type What it is video_urlrequiredstring Video file or finished HLS playlist (.m3u8) to extract frames from. modestring timeline (default), timestamps, single, batch, first, middle or last. timestampsnumber[] Exact positions in seconds. Selects timestamps mode. positionstring For single mode: first, middle or last. Default: middle. frame_intervalnumber Batch only: take every Nth frame (1-300, not seconds). Default: 12. max_framesnumber Timeline 1-24 (default 12); batch 1-500 (default 100). output_formatstring Batch or timestamps: png (default), jpg or webp. qualitynumber Batch or timestamps with jpg or webp: 1-100. Default: 95. start_timenumber Timeline only: start of the section. end_timenumber Timeline only: end of the section. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. generate_html_video
Generate HTML VideoChanges dataParameters (12)
Name Type What it is promptrequiredstring The brief for a new video, or what to change when cloud_project_id is given. Name the user's own assets and their roles; ask the user for media the brief needs and you do not have. cloud_project_idstring Project ID from an earlier result. Pass it to change that video; omit it to start a new one. durationnumber Seconds, up to 300. Default: 5, or the project's. widthinteger Even number, 2-3840. Default: 1920, or the project's. heightinteger Even number, 2-3840. Default: 1080, or the project's. fpsinteger 24, 30 or 60. Default: 30, or the project's. format"mp4" | "webm" | "mov" Default: mp4, or the project's. effort"light" | "medium" | "high" light for simple changes and basic videos, medium (default) for most, high for complex or polished videos; higher is slower and may cost more. A change keeps the project's effort unless given. media_budget_creditsinteger Most credits the agent may spend on generated media this turn (default 200; 0 for none). reference_assetsobject[] Files the user supplied or selected (logos, images, audio). Images are shown to the agent; every file can be used in the video. project_urlstring Publicly downloadable HTTPS URL of a HyperFrames project ZIP to start a new video from, such as a ZIP URL returned by search_examples({ id }). project_asset_idstring Library ZIP or project asset to start a new video from. It is unpacked as the starting project. get_film_project
Get Film ProjectRead-onlyParameters (1)
Name Type What it is idrequiredstring Film project id from create_film_project or list_film_projects. get_theme
Get themeRead-onlyParameters (1)
Name Type What it is namestring Theme name. If omitted, returns the default theme. import_media
Upload mediaRead-onlyParameters (2)
Name Type What it is purpose"media" | "voice_clone" voice_clone: one private voice sample; the card returns audio_asset_id for clone_voice instead of a public URL. contextstring What the files are for; shown on the card. increase_frame_rate
Increase frame rateChanges dataParameters (5)
Name Type What it is video_urlrequiredstring Finished video to retime. modelstring Frame interpolation model ID. Default: frame-interpolation/topaz-apollo. extrasobject Model-specific parameters as listed by get_model_params. They are merged with the top-level fields. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. isolate_audio
Isolate audioChanges dataParameters (3)
Name Type What it is audio_urlrequiredstring Audio to clean (mp3, wav, m4a, ogg, aac), or a video whose soundtrack to clean. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. lip_sync_video
Lip sync videoChanges dataParameters (7)
Name Type What it is video_urlrequiredstring The existing video whose mouths should move. audio_urlstring The voice, which replaces the video's audio. Leave it out when the video already has the voice in it. promptstring Who says what, in plain words, for example "the first voice is the owl, the second is the squirrel". Used with model sync-3; leave it out when the speakers are obvious. audio_start_secondsnumber Seconds into the video where the voice in audio_url starts, when it is shorter than the video and belongs at a set moment. Only that stretch is lip-synced; the video's sound plays around and under the voice. modelstring pixverse (default) or sync-3. Leave it out unless the user names a model. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. list_budget_projects
List Budget ProjectsRead-onlyParameters (2)
Name Type What it is budget_projectstring Optional: one project's slug, id, or name, to list its latest generations. include_archivedboolean Include archived projects in the list. list_characters
List CharactersRead-onlyParameters (0)
No parameters.
list_film_projects
List Film ProjectsRead-onlyParameters (0)
No parameters.
list_review_requests
List review requestsRead-onlyParameters (0)
No parameters.
list_speech_pronunciations
List speech pronunciationsRead-onlyParameters (4)
Name Type What it is family"elevenlabs" | "fish" | "google" | "minimax" | "chatterbox" | "kokoro" | "orpheus" Model family, shared across versions and voices: elevenlabs covers v2 and v4, google covers both Gemini TTS models, fish covers Fish Drama 3. languagestring Language code, for example he or en. Regional codes share their base-language rules. limitinteger Entries to return. offsetinteger Entries to skip: the nextOffset of the previous page. list_templates
List TemplatesRead-onlyParameters (2)
Name Type What it is include_htmlboolean Also return each full template. Default: false. name_containsstring Case-insensitive filter on name and description. list_themes
List themesRead-onlyParameters (0)
No parameters.
merge_media
Merge MediaChanges dataParameters (12)
Name Type What it is video_urlsstring[] 2 to 25 videos in playback order. Pass this or audio_urls. video_fit"auto" | "crop" | "pad" | "strict" How clips of a different shape fit the canvas. Default: auto. canvas_video_indexinteger Zero-based index of the video that sets the canvas. Default: 0. pad_color"black" | "white" | "gray" Bar color for video_fit=pad. Default: black. target_fpsnumber Output frame rate, 1 to 60. Default: the lowest input frame rate. audio_urlsstring[] Audio files in order. Only the first 5 are joined per job. audio_format"mp3" | "m4a" | "wav" Audio output. Default: wav for WAV-only inputs, m4a for AAC-only inputs, otherwise mp3. crossfade_secondsnumber Audio: overlap each adjacent pair, 0 to 30 seconds. normalize_loudnessboolean Audio: match each input's loudness before joining. target_lufsnumber Audio: target loudness, -30 to -5. Default: -18. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. open_studio
Open StudioRead-onlyParameters (1)
Name Type What it is page"home" | "explore" | "library" | "characters" | "voices" | "approvals" | "billing" | "settings" Page to open: home, explore (examples to start from), library, characters, voices (saved designed and cloned voices), approvals, billing or settings. Omit for Home. overlay_images
Overlay ImagesChanges dataParameters (4)
Name Type What it is video_urlrequiredstring Video to put the images on. overlaysrequiredobject[] Later entries appear above earlier ones. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. rate_asset
Rate assetChanges dataParameters (4)
Name Type What it is asset_idrequiredstring The asset being rated. rating"up" | "down" | null Thumbs up, thumbs down, or null to clear. favoriteboolean Keep it in Favorites. notestring Optional feedback note. render_html_image
Render HTML ImageChanges dataParameters (9)
Name Type What it is htmlrequiredstring HTML to render. Write it as for a real browser. Use {{token}} for inline_images. widthinteger Output width in pixels. Default: 1200. heightinteger Output height in pixels. Default: 630. inline_imagesobject[] Images or fonts to embed. Every {{token}} needs a matching entry. transparent_backgroundboolean Make unused pixels transparent in the PNG. namestring Optional asset name for the saved file. tagsstring[] Optional tags for the saved file. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. render_html_video
Render HTML VideoChanges dataParameters (12)
Name Type What it is htmlstring The composition. Pass this, project_url or project_asset_id. project_urlstring Publicly downloadable HTTPS URL of a static project ZIP, e.g. a ZIP URL from search_examples({ id }). project_asset_idstring A ZIP in the Library instead of html: a project asset (get_upload_url type project, recommended) or an ordinary zip asset. Static projects only: index.html or a built dist/index.html. durationnumber Seconds, up to 300. Default: 5. fpsnumber 24, 30 or 60; others are rounded to the nearest. Default: 30. widthinteger Even number, 2-3840. Default: 1920. heightinteger Even number, 2-3840. Default: 1080. format"mp4" | "webm" | "mov" Default: mp4. namestring Optional asset name for the saved file. tagsstring[] Optional tags for the saved file. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. render_template
Render TemplateChanges dataParameters (10)
Name Type What it is template_idstring Template ID. Pass this or template_name. template_namestring Template name, ignoring case. Pass this or template_id. rendersrequiredobject[] Variants to render, 1 to 4. size_namesstring[] Named sizes to render, or ["all"]. width_overrideinteger Override the width of a single-size render. height_overrideinteger Override the height of a single-size render. namestring Optional asset name for the saved file. tagsstring[] Optional tags for the saved file. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. scale_video
Scale VideoChanges dataParameters (7)
Name Type What it is video_urlrequiredstring Video to scale. widthinteger Target width, 2-7680, even. heightinteger Target height, 2-4320, even. mode"stretch" | "pad" | "crop" stretch distorts to fit, pad letterboxes, crop center-crops. Default: stretch. pad_color"black" | "white" | "red" | "green" | "blue" | "gray" Bar color for mode=pad. Default: black. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. search_examples
Search ExamplesRead-onlyParameters (8)
Name Type What it is querystring What the user wants to make. output_type"image" | "video" | "audio" Only this media type. model_idstring Only examples for this model. render_type"html_video" html_video finds runnable HyperFrames examples. tagsstring[] Every result must have all of these. limitinteger Results per page. Default: 30. cursorstring From a previous page. idstring Load one example in full; other filters are ignored. search_voices
Search voicesRead-onlyParameters (7)
Name Type What it is querystring A few words on how the voice should sound or what it is for, such as warm calm narrator or energetic ad. Put language, gender, and accent in their own fields. modelstring Exact speech model ID, such as speech/elevenlabs-v4, to return only voices that work with it. Omit to search every speech model. gender"male" | "female" Voice gender. languagestring Language the voice will speak: a code or English name, such as es, en-AU, or Spanish. Multilingual voices also match. accentstring Accent or region, such as australian, british, or southern. limitinteger Voices to return. offsetinteger Voices to skip, to get the next page of the same search. set_review_mode
Set review modeChanges dataParameters (1)
Name Type What it is moderequired"review" | "auto" review or auto. show_budget_project
Show Budget ProjectRead-onlyParameters (2)
Name Type What it is budget_projectstring Project to open: its slug, or an id or name. Omit to show all projects. tab"generations" | "media" | "approvals" | "projects" Tab to open: generations (default, one row per charged generation), media, approvals or projects (all projects). stitch_clips
Stitch ClipsChanges dataParameters (5)
Name Type What it is clipsrequiredobject[] Clips in playback order. audioobject Optional audio to mix over the clips. content_fit"contain" | "cover" Default: contain. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. swap_video_characters
Swap Video CharactersChanges dataParameters (7)
Name Type What it is video_urlstring Direct video file URL, at least 4 seconds. Up to 30 seconds is used; longer sources are cut. Import social post links with upload_asset first. image_urlsstring[] Reference photos. One photo may contain several replacement people; describe which person in each photo replaces which person in the source video. character_idsstring[] Saved Characters to swap in; their images count toward the 8 references. instructionsstring Describe the edit in your own words: who in the video becomes whom, including left/right or clothing descriptions within group photos, who stays unchanged, and any requested face, hair, clothing or other visual changes. Explicit instructions guide the previews and video; extra photos do not require extra replacements. resolution"480p" | "720p" | "1080p" 480p (default) renders a native draft that can be rendered at 1080p later from the card; 720p or 1080p render that resolution directly. video_model"video/seedance-2.5" | "video/wan-3.0" | "video/kling-4.0-omni" video/seedance-2.5 (default, Pro): up to 30 seconds of input video and 8 combined references. video/wan-3.0 and video/kling-4.0-omni are not supported for swaps and return an error. reviewboolean Pass true only when the user asks to check the previews before rendering: it keeps the Review step even if the account renders in Auto mode. Omit it otherwise; the account setting applies. transcribe
TranscribeChanges dataParameters (6)
Name Type What it is audio_urlstring A direct audio file, or a video whose soundtrack to transcribe. Pass this or video_url. video_urlstring A direct video file, supported social post or public Google Drive link. Social sources try transcript retrieval first, then download and transcribe the audio if necessary. Pass this or audio_url. language_codestring ISO-639 code (e.g. "en", "es"; "en-GB" is read as "en"). Social sources: captions in another language are skipped for the audio fallback. Speech to text detects the language itself; a transcript plainly in another language comes with a warning. max_durationinteger Social sources: the longest full source in seconds, not a trim range. Default 7200; longer sources are rejected. The complete bounded quote must fit the balance. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. trim_video
Trim VideoChanges dataParameters (8)
Name Type What it is video_urlstring Video to trim. Pass this or audio_url, not both. audio_urlstring Audio to trim, or a video whose soundtrack should be cut. Returns audio only. Pass this or video_url, not both. start_timenumber Start in seconds. Default: 1. end_timenumber End in seconds. Takes precedence over duration. durationnumber Seconds from start_time. Default: 2. Ignored if end_time is given. audio_format"mp3" | "m4a" | "wav" Audio output format when audio_url is given. Default: mp3. force_newboolean An identical request submitted in the last 10 minutes returns that job without charging again; true generates it again. budget_projectstring Optional budget project this generation's cost counts toward, for tracking spend per client or job. Pass its slug; an id or the exact display name also works. Create one with create_budget_project and find slugs with list_budget_projects. An unknown project never stops the generation: it runs without one, and the result warns how to add it with assign_to_budget_project. update_budget_project
Update Budget ProjectDeletesParameters (5)
Name Type What it is budget_projectrequiredstring The budget project's slug, e.g. papercraft-tarot. An id or the exact display name also works. namestring New display name. slugstring New slug. notesstring | null New notes; null clears them. archivedboolean true archives the project, false restores it. update_character
Update CharacterDeletesParameters (11)
Name Type What it is idrequiredstring ID of the Character to update. titlestring New name. descriptionstring Appearance, personality, role: added to prompts at generation time. image_urlstring Reference portrait or character sheet URL: the visual anchor for generate_image and generate_video. profile_image_urlstring Small browsing portrait, stored separately from image_url. If omitted when adding or changing a sheet, a profile portrait is generated asynchronously with normal image credits. For a portrait-only Character, pass the portrait URL as both image_url and profile_image_url. voice_modelstring Speech model of an existing public voice, such as speech/elevenlabs-v4. voice_idstring Exact public voice ID for voice_model, from get_model_params. This does not create a new voice. voice_namestring Optional display name for the saved public voice. voice_preview_idstring preview_id of a design_voice audition to save as a voice of this Character. It becomes a private Fish voice that speaks with speech/fish-drama-3. voice_languagestring | null Optional voice preview language code, such as en, he or es. Null clears it. With a saved voice, setting or changing language generates one short localized preview asynchronously using normal speech credits. Omit to keep the current value; no language means no automatic preview. replace_voice_option_idstring With voice_preview_id, replace this saved designed voice option instead of adding another. The old provider voice is deleted. Only on the user's explicit request. update_film_project
Update Film ProjectDeletesParameters (13)
Name Type What it is idrequiredstring Film project id from create_film_project or list_film_projects. namestring New project title. briefstring New creative brief. loglinestring The film's one-line premise. status"drafting" | "script_ok" | "storyboard_ok" | "rendering" | "preview_ok" | "final" Approval stage; moves forward one stage per call. character_idsstring[] Character ids for the cast. theme_idstring Brand theme id. target_duration_snumber Target total film length in seconds; fractional values are rounded. audio_urlstring Full narration track URL assembled_urlstring Final stitched film URL shotsobject[] Full validated shot list; replaces the existing list. patch_shotsobject[] Validated partial shot updates merged by stable id. budget_projectstring | null Tie this film to a budget project (slug, id or name); null removes it. update_template
Update TemplateDeletesParameters (13)
Name Type What it is template_idstring Template ID. Pass this or template_name. template_namestring Template name, ignoring case. Pass this or template_id. namestring New name. descriptionstring New description. htmlstring html kind: replacement markup. widthinteger html kind: new width. heightinteger html kind: new height. sizesobject[] Pass [] to clear and use width/height. parametersobject[] Replacement parameter list. promptstring generative kind: replacement prompt. recommended_modelstring generative kind: new model ID. model_paramsobject generative kind: replacement model parameters. reference_asset_idsstring[] generative kind: replacement reference assets. update_theme
Update themeDeletesParameters (6)
Name Type What it is namestring Theme name. Creates the theme if it doesn't exist. Required unless interactive is true. dataobject Theme data: colors, fonts, logos, etc. Merged with existing data by default. imagesstring[] Brand reference images for this theme. Replaces the entire images array when provided. The first image is saved as the thumbnail. overrideboolean If true, replaces the entire theme data. Default false: top-level keys are replaced and missing keys are preserved. Only applies to data. set_defaultboolean If true, makes this theme the default. interactiveboolean When true, open the visual theme editor in the chat instead of saving. The user fills in the form and saves there; name is optional. Use this whenever the user says "edit my theme", "create a theme" or "help me set up a brand". update_video_workflow
Update Video WorkflowDeletesParameters (8)
Name Type What it is session_idrequiredstring The session ID from the workflow’s result, for example a swap_video_characters result. expected_versionrequiredinteger The version from the latest result for this session. A stale version is refused so a newer change is never overwritten; check_job with the session ID returns the current one. preview_indexinteger Which preview to change, counting from 0 in the result’s previews list. changesstring What to change in that preview, in plain words. source"previous" | "original" Refine the previous preview, or restart from the original source frame to fix pose, anatomy, crop or background drift. Defaults to previous. scope"frame" | "all" Update one frame (default) or redo every preview from the original source frames. source_video_job_idstring Accepted for compatibility: sources are fitted automatically, so it is never needed. video_model"video/kling-4.0-omni" | "video/wan-3.0" Not supported: only Seedance 2.5 renders swaps, so these values return an error.