add_influencer_look
ChatGPTUse this when the user wants a new outfit or style for one of their AI Influencers ("give franck a gym look"). Takes a short name and what they wear; the card shows the look when it is drawn.
Overview
Turn one photo into a short video: put yourself into trending templates like Hotel Lobby, Couture Crowns, Fan Fits, and NDA Glow Up, see a preview of you in it, then make the full video. Or create and refine faceless videos, UGC-style ads, and explainers from an idea, in the same GoFaceless account.
Tools
add_reference
ChatGPTAttach one organization-owned media ID or inspect one safe HTTPS URL. Use the CLI media command first for a local file.
approve_video
ChatGPTUse this only after the user said yes to a draft's exact price ("go ahead", "make it", "yes for 300 credits"). Makes that draft version and spends at most the credits they saw. If the price or the plan changed, no credits are spent and the new version or price comes back to show them. Approving the same version again returns the same video.
create_ai_influencer
ChatGPTUse this when the user wants their own AI Influencer (a reusable character with their face, a voice, and outfit looks) made from their photo ("make an AI influencer of me"). Uses the photo as the face, designs a voice from their words, and starts drawing the signature look; the card shows it when ready. Confirm the photo is of the user or someone who agreed.
decide_pending_action
ChatGPTDecide only the exact pending action and state version returned by get_video_session. Approval may allow paid generation to continue.
export_video
ChatGPTUse this when the user wants the final file of a finished chat video session: "give me the final MP4", "export it", "I want the video file". Delivers the current state-bound final MP4 as a download link when the account's plan includes downloads; otherwise it says so, and the video stays watchable on GoFaceless. This never publishes to a social network. Do not use for a public watch link (use share_video), for a draft (use render_video_preview), or for template videos (use get_template_video).
get_template_preview
ChatGPTRead the preview still that prepare_template_video started: drawing, ready with its image, or failed. Shows the preview in a card; call it once when the user asks to see the preview, never to poll.
get_template_video
ChatGPTUse this when the user asks whether their GoFaceless template video is ready, or for its link. Reads a template video's progress, finished video link, and GoFaceless link. Omit job_id to read the latest template video of this account, so this also works in a new chat. Shows the video in a card that plays it when ready and updates itself: call it once per user question, never to poll, and never start the video again while it is running.
get_video
ChatGPTUse this when the user asks whether a GoFaceless video is ready, how far along it is, or what a draft says ("is my video done?"). Reads one draft or video: its plan and price, or its progress and result. The card already follows a video being made, so never call this just to poll.
get_video_capabilities
ChatGPTDiscover account, format, reference, edit, host-fallback, and phase-specific GoFaceless capabilities. Call once before starting a session or when a requested action is unavailable, not during routine polling.
get_video_session
ChatGPTMonitor one existing video session. Obey poll_after_ms, keep polling the same session while the turn remains active, and never start a duplicate while waiting. Returns compact text-only status so routine polls do not create duplicate UI cards. To show, play, or preview a video for the user, use show_video instead.
list_ads
ChatGPTUse this when the user asks to see their ads or whether their ads are ready ("show my ads", "are my Instagram ads done?"). With batch_id, shows that set of ads; without it, the latest ads. Shows the ad images and their video ads in a card that updates itself, so call it once per question, never to poll. Do not use for chat session videos or template videos (list_video_sessions, get_template_video).
list_ai_influencers
ChatGPTUse this when the user wants their saved AI Influencer (a character they created on GoFaceless, with a face, a voice, and looks) in a template, or asks which influencers or looks they have. Lists each influencer with their ready looks. To star one in a template, pass the look IDs to prepare_template_video as influencer_look_ids instead of attaching photos.
list_influencer_photos
ChatGPTUse this when the user asks to see the content photos of one of their AI Influencers, or before making a pack to learn its price. Shows the photos with their captions in a card, and gives the credits per photo and how many photos the account can still make today.
list_series
ChatGPTUse this when the user wants to carry on, check, or pick one of their existing animated series and its series_id is not in this chat: "continue my cartoon series", "what happened to my pizza series?", "write the next episode of my show". Lists the account's series, most recently active first, with parts made, the latest part's status, and whether the next part can be written. Spends nothing. Do not use to start a new series (write_series) or to list finished videos (list_video_sessions).
list_video_sessions
ChatGPTUse this when the user asks to see their GoFaceless videos ("show my latest videos", "my library"): shows them as one grid card in which every video plays. Set order to oldest for "my first videos" and limit to the number asked for. That card is the whole answer; do not open each video with show_video.
list_video_templates
ChatGPTUse this when the user wants to put themselves, a friend, or a partner into a viral video, trend, or template from a photo: "put me in this trend", "face swap me into a music video", "make me a glow-up video", "turn my selfie into a viral video", and has not picked a template yet, or asks which templates there are. If the user already names a template (Hotel Lobby, Who's Driving, Couture Crowns, Abracadabra, Fan Fits, NDA Glow Up, ...), call prepare_template_video directly instead. Templates sit in the same categories as gofaceless.ai/templates (Anime, Fashion, Culture, Cartoon). When the user asks for a kind of template ("anime templates", "fashion", "something for a couple", "glow-ups"), pass their words as query, or pass templates with the ones you mean, so the card shows only those. Lists the templates, who is in each, how many photos each takes, and their options. Not for editing a single still picture; the result is a short video.
list_videos
ChatGPTUse this when the user asks what GoFaceless videos they have ("my videos", "what did I make?"), or to find the ID of a video or draft they name. Lists them newest first with their status.
make_ad_video
ChatGPTUse this only after the user approved the exact price from quote_ad_video ("yes, make the video ad", "go ahead with that video"). Makes one short video ad from a finished static ad (the same product, words, and brand, ending on the ad; a promo is 16:9 and ends on the product name) and charges the approved credits once; a retry with the same arguments never charges twice. It takes several minutes; list_ads shows it when ready. Do not use for chat session videos or template videos.
make_influencer_photo_pack
ChatGPTUse this when the user wants ready-to-post photos of one of their AI Influencers ("make Maya a gym photo pack", "photos of franck for Instagram"). Each photo puts them in one of their looks somewhere new, with a caption in their voice. It costs credits: get credits_per_photo from list_influencer_photos, tell the user the price of the pack (count x credits_per_photo), and call this only after they approve it, with approved_credits set to that price. Photos that cannot be drawn are refunded. The card shows the photos as they are drawn (a minute or two each).
make_series_part
ChatGPTUse this only after the user approved the exact price of a series part shown with its script ("yes, make it", "make episode 1"). Spends that many credits to make the part as an animated video, at the approved price and for the exact draft that was priced; it is refused if the price went up or the script changed since. Returns a job ID and shows the part's progress and video in a card. Do not use without that approval, or for template or faceless videos.
make_static_ads
ChatGPTUse this only after the user approved the exact price from quote_ads ("yes, make them", "go ahead with the 8 ads"), with the arguments quote_ads gave. Makes finished static ads for each product (feed 4:5, square 1:1, and story 9:16 sizes) in the brand's colours, with words taken from its own page, and spends the approved credits once; a retry with the same client_request_id never spends them twice. Returns right away: writing the ads takes a few minutes, and the card shows each ad as it is drawn. Do not use for template videos or faceless video sessions (start_video_session).
make_video_variant
ChatGPTUse this when the user wants a finished GoFaceless video in another shape or length ("make a square version", "a vertical one for TikTok", "a 15-second cut"). Makes a new video from the same setup with that one change; it costs credits like a new video, so confirm with the user first. Template videos keep their own format.
manage_cast
ChatGPTUse this when the user asks who is in one of their shows (an AI Series), or wants one of their AI Characters added to or removed from it ("put Maya in my bakery series"). Lists their shows, shows one show's cast, or adds or removes a character, so every later part keeps the same cast.
plan_video
ChatGPTUse this whenever the user wants a new GoFaceless video, whatever kind: themselves in a template from their photo ("put me in Hotel Lobby", "make a glow-up video of me"), their face in a pasted video link, a faceless or narrated video from an idea ("a 30-second video about saving money"), a whiteboard explainer, an ad, an AI Series show or its next part, several videos at once, or another shape or length of a video they have. Returns a free draft with its credit price, or what it still needs (photos, or the page that makes that kind of video). Show the user the draft and price; nothing is made and no credits are spent until approve_video.
prepare_template_video
ChatGPTUse this whenever the user attaches a photo and asks to be put into a template by name (Hotel Lobby, Neptune Square, Who's Driving, Couture Crowns, Abracadabra, Fan Fits, Strut, NDA Glow Up, Smooth Operator, One Piece Gear 5, Super Saiyan, Sage Mode, Domain Expansion, Bankai, Sun Breathing, Killua Godspeed, Gon Adult), even without saying "video" or "GoFaceless" ("put me in the Couture Crowns template with this photo"). These are GoFaceless video templates, not pictures to draw: never answer such a request with image generation or by drawing it yourself. Call this first; do not list templates first. It shows them in the template as a preview still and gives the exact price of the full video. Adds the user's photos or AI Influencer looks (one per person to replace) to a template, start a free preview still of them in it (free on every plan), and get the exact credit price without spending anything. For "my last photo" or "the same photo" with no photo IDs in this chat, set recent_photos instead of asking for the photo again. Confirm the user owns or has permission to use each photo. Show the price and ask the user to approve it before start_template_video.
quote_ad_video
ChatGPTUse this when the user wants one of their finished ads as a short video ad ("a video ad for this product", "turn this ad into a video", "a presenter video of this ad", "a promo video for my site"). Needs an ad_id from list_ads or make_static_ads; a promo of the whole site takes the site_id from read_brand_site instead. Gives the exact credit price of that video (cut: the ad filmed; presenter: a creator says the ad's line; motion: the product shot moves with a camera preset; promo: a 16:9 film from the brand's site, one word on each music beat over its own photos, ending on the product name) and lists the motion presets; spends nothing. Show the price and ask the user to approve it before make_ad_video. Do not use for a video from an idea or script (start_video_session) or for putting the user into a template.
quote_ads
ChatGPTUse this before making static image ads (Ads Studio): right after read_brand_site with its site_id and the product_ids the user picked, or for a product the user describes with its photo attached ("Instagram ads for my product"). Gives the exact credit price and the account's credits; spends nothing. Set videos_per_product when the user also wants video ads with the set ("ads and a video ad for this product"). Show the price and ask the user to approve it before make_static_ads. Do not use for template videos or chat video sessions.
quote_video_plan
ChatGPTUse this after writing a plan of several videos and before asking the user to approve it: gives each video's credit price, the total, and whether the account has enough. Spends nothing.
read_brand_site
ChatGPTUse this when the user wants ads for their shop, brand, or product and gives its website or a product page link ("make ads for my shop https://…", "Instagram ads for my product https://…", "Facebook ads for this store"). Reads the public site into the brand (name, colours, voice) and its products with their photos, and shows the products in a card. Spends nothing. Reading takes one or two minutes, so it may answer that it is still reading: the card shows the products when they are found, and calling it again with the site_id it gave checks on the read. Next, call quote_ads with its site_id and the products the user picks. Do not use for general web browsing, for putting the user into a video template, or for a faceless video about a topic (start_video_session).
render_video_preview
ChatGPTQueue a state-bound scene, range, or full preview through the existing GoFaceless render gates.
revise_series_script
ChatGPTUse this when the user wants to change the script written by write_series or write_next_series_part before it is made: "make it funnier", "change line 3 to…", "make Maya the villain", "make it 30 seconds", "end the story here". Takes an instruction, exact new words for numbered lines, or the edited script. Spends no credits; returns the new script and its new price. Do not use to change a part that is already made, or for template or faceless videos.
revise_video_plan
ChatGPTUse this when the user wants to change a draft from plan_video before it is made ("make the opening funnier", "make it 15 seconds", "use a male voice", "use my other photo"). Saves a new version and returns it with its price, free. Show the new version and price before approve_video.
send_video_instruction
ChatGPTSend natural-language creative or edit intent to the existing durable GoFaceless session agent.
share_video
ChatGPTUse this when the user wants a link to share a finished GoFaceless video ("share my video", "give me a link to send"). Creates a public page anyone with the link can watch, or returns the existing one. Only for finished videos; the owner can turn the link off in GoFaceless.
show_account
ChatGPTUse this when the user asks about their GoFaceless credits, plan, renewal, free previews, or what they spent recently ("how many credits do I have?"). Shows it in a card.
show_video
ChatGPTUse this when the user wants to see or watch one of their GoFaceless videos ("show me my bakery video", "play my last video"). Opens the video in a card that plays it once it is ready, or shows its progress until then. Omit session_id for the latest video. Only for one video the user asked to watch: never call it for each video after list_video_sessions, whose grid already plays them. Use get_video_session only to check status.
start_template_video
ChatGPTUse this only after the user approved the exact price from prepare_template_video. Spend credits to make the template video, using the photo IDs from prepare_template_video and the exact price the user approved. Returns a job ID; monitor it with get_template_video.
start_video_batch
ChatGPTUse this after the user approved a plan of several videos ("make all five", "plan this week's videos and make them"). Price the plan with quote_video_plan first and show the total; after the user approves it, call this with approved_credits set to that total. Starts up to five faceless videos at once, in plan order, each only if it fits within approved_credits and the credits left; the rest are reported as not started. Shows them together in a grid that updates until each is ready.
start_video_session
ChatGPTUse this when the user wants a new short video made from an idea, script, topic, product, or reference (faceless explainers, UGC-style ads, listicles, AI creator videos). Do not use when they want their own face or photo put into a viral trend or template; use the template tools for that. Starts one durable video session. Defaults to autonomous execution; use controlled mode only when the user explicitly requests review checkpoints. Returns immediately with an operation and session ID.
undo_last_video_change
ChatGPTRestore a captured pre-change video state. Read the current session first and provide its exact state version.
watch_video
ChatGPTUsed by the GoFaceless card to follow a video while it is made.
watch_video_session
ChatGPTUsed only by the GoFaceless video card to follow a session while it is made: waits up to wait_seconds for a change, then returns the session. Not for the model; use get_video_session.
write_next_series_part
ChatGPTUse this when a series part is finished and the user wants the next episode: "write part 2", "what happens next?", "next episode where they get caught". Writes the next part's script from the season plan for free, and returns it with its price. Do not use before the latest part is finished, to change the current script (revise_series_script), or to start a new series (write_series).
write_series
ChatGPTUse this when the user wants a cartoon or animated series, a story told in episodes, or a skit or short drama acted out by two characters: "make a cartoon series about my cat and my dog", "episode 1 of my story about a haunted bakery", "an animated skit with two characters arguing about pineapple pizza", "a short drama series". Plans the season and writes the Part 1 script for free, returns the script to show the user, and the credit price of making Part 1 as a video. Nothing is spent. Do not use for faceless explainers, ads, or narrated videos about a topic (start_video_session), or for putting the user's own face or photo into a template or trend (the template tools).
Capabilities
InteractiveWrites
Example Prompts
Click any prompt to copy it.
