add_scene
ChatGPTAdd a scene during script review (before approve_script) with its narration. position is 1-based; omit to append.
approve_script
ChatGPTApprove the reviewed script and START RENDERING (10-15 minutes). Fabeo then writes each clip's H3 prompt from the script and renders it. Spends credits, up to the returned estimate -- show the user estimate_cost first and get their OK.
check_audio_upload
ChatGPTWhether the user has uploaded their recording on a get_audio_upload_page link yet, and its length.
check_reference_upload
ChatGPTHow many photos the user has uploaded on a get_reference_upload_page link so far.
create_from_template
ChatGPTCopy a template from list_templates (pass its template_job_id) into the user's account. The copy lands at script review (free) and is word-for-word the template: its narration and visuals still describe the template's own subject. Before approve_script, rewrite every scene for the user's product, brand and call to action with edit_scene (the result lists the scenes), keeping the template's structure and style; scenes can be deleted but not added. Reference photos can't be applied to a copy -- for a video that shows the user's own person or product photos, use create_video with a topic modelled on the template. single_scene=True renders the template's first scene unchanged, immediately, and spends credits.
create_video
ChatGPTStart a new video from a topic (10-1500 chars: subject, style, message, call to action). Multi-scene videos first get a script for review (free); nothing is rendered or billed until approve_script. single_scene=True makes one short clip (<=11s) that renders immediately and spends credits right away. aspect_ratio: '9:16' (vertical) or '16:9'. video_type: one of Fabeo's seven styles (see get_started); its style is written into the brief. reference_image_keys (from upload_reference_image / get_reference_upload_url) and/or reference_upload_id (from get_reference_upload_page, after the user uploaded): up to 5 images in total; every scene is then rendered to match them (the person's face/outfit, the product), so describe the person in the topic consistently with those images. product_id (from list_catalog): the video is for that product -- its description goes into the brief, and catalog_image_keys (from get_product) adds its stored photos as references.
create_video_from_audio
ChatGPTMake a video from the user's own voice recording plus reference photos: one scene, exactly as long as the recording (up to 60s), showing the person in the photos speaking it, lip-synced -- in the recording's own words, voice, language and accent. Pass the recording as audio_key (upload_audio / get_audio_upload_url) or audio_upload_id (get_audio_upload_page), and the photos as reference_image_keys and/or reference_upload_id (strongly recommended: without them Fabeo invents the person). topic (optional) describes the look, setting and style. Renders immediately and spends credits (~1 per second of the recording) -- tell the user the estimate and get their OK first; pass an idempotency_key.
delete_scene
ChatGPTDelete a scene during script review (before approve_script). Irreversible.
delete_video
ChatGPTPermanently delete a video job with all its scenes and files. Requires confirm=true; ask the user first.
edit_scene
ChatGPTChange one scene; pass only what changes. Nothing is re-rendered here: the result's regenerate_mode tells you which regenerate_scene mode applies it (none during script review, where approve_script renders everything anyway). - narration: the exact words spoken. Use it to fix mispronounced, wrong or garbled words (write names/numbers the way they should be said). On a rendered video Fabeo re-writes the scene's prompts with the new line on regenerate_scene(mode='narration'). If the voice is still wrong after that (pronunciation, a regional language or accent, gibberish), offer the user to record the line themselves: set_scene_audio. - visual_description: what the scene shows (setting, action, outfit, camera). Also re-written into the prompts on regenerate_scene(mode='narration'). - positive_prompt (rendered scenes only; clip clip_index, see get_scene_prompts): write the clip's MiniMax H3 prompt yourself for exact control. It drives BOTH picture and sound, so it must quote the narration word for word in its audio part (the whole line for a one-clip scene, this clip's consecutive part otherwise) -- it is rejected if it doesn't; to change the words pass narration too. H3 style: one continuous take (no cuts or montage), concrete subject/appearance/outfit/setting/camera motion, no on-screen text; in reference-image scenes open with '<Subject 1> is ... in <Picture 1>'. A hand-written prompt is kept as written (Fabeo won't overwrite it). - keyframe (rendered scenes only): 'keep' (default); 'variant' to try a different key frame for the clip; 'upload' to open the clip on the user's own image -- upload it first with get_start_frame_upload_url (the clip then animates from that exact frame, so describe motion from it in positive_prompt). - duration_seconds: the scene's length (2-15s per clip). The clip renders exactly this long, so shorten it to cut music, silence or garbled speech after the voiceover ends; a length shorter than the narration needs cuts the voiceover (the result warns). Shorter scenes also cost fewer credits. Applied by regenerate_scene(mode='clips') on a rendered video. - start_frame_image_key / end_frame_image_key: make the scene open on, and/or land on, the user's own pictures. Works during script review (before approve_script) and on a rendered video. Get the key first like a reference image (upload_reference_image, get_reference_upload_url, or get_reference_upload_page + check_reference_upload; this needs the user's rights confirmation) -- not get_start_frame_upload_url, which is per rendered clip. The start frame is the scene's first clip's opening picture; the end frame is where its last clip should finish. They must be two different images; they are separate from the scene's reference images (the people/products shown). Describe the motion between them in positive_prompt. Check them in get_video / get_scene_prompts (start_frame, end_frame); remove_start_frame / remove_end_frame undo. On a rendered video apply with regenerate_scene(mode='images').
edit_scene_narration
ChatGPTChange one scene's spoken line. Same as edit_scene(narration=...), which is preferred: on a rendered video, apply it with regenerate_scene(mode='narration'), which re-writes the scene's prompts so the clip speaks the new line.
estimate_cost
ChatGPTEstimate the maximum credits a render would cost (only clip generation is billed, ~1 credit per rendered second; failed renders are refunded in full). For regenerate_scene pass the chosen scenes' total duration. Always share this with the user before approving a script or regenerating.
get_account
ChatGPTShow the account's subscription plan and remaining credits, how much this connection may still spend (its safety limit, separate from the credits), and links to the account's plan and billing details. Check it before a render if credits may be short.
get_audio_upload_page
ChatGPTLet the user upload their own voice recording from their device: returns a one-hour upload_page_url to give them (one MP3/WAV/M4A/OGG file, up to 60s and 4 MB; they confirm they hold the rights to it there). Use in chat apps, which can't upload files. When they say they're done, call check_audio_upload, then create_video_from_audio or set_scene_audio with the audio_upload_id.
get_audio_upload_url
ChatGPTFor clients that can make HTTP requests (e.g. Claude Code): a one-hour presigned upload for one voice recording (mp3/wav/m4a/ogg, up to 60s). Upload it with an HTTP multipart POST (every upload.fields entry as a form field, then the file as 'file'), then pass the returned audio_key to create_video_from_audio or set_scene_audio. rights_confirmed must be true: first tell the user they must own the recording or hold the rights to it (including the speaker's consent) and that Fabeo isn't responsible for infringement, and get their OK. Chat apps: use get_audio_upload_page or upload_audio.
get_job_status
ChatGPTCheck a job's progress -- share it with the user while they wait: stage ('step 4 of 5: rendering video clips (3 of 6 scenes rendered)'), progress_percent (0-100) and per-stage statuses, plus a next_action hint. A render moves SCRIPT -> AUDIO -> KEY_FRAMES -> CLIP_GENERATION -> VIDEO_STITCH. Follow next_action rather than 'status' alone: while regenerate_scene runs, status stays COMPLETED and next_action says Regenerating (video_url is still the previous version until then). On a failure it gives the reason and a credits entry with what the job was actually charged and refunded.
get_product
ChatGPTOne product: title, description, extra context, its brand and client, and every stored photo of the product, brand and client (image_key, url, source) -- show them to the user and pass the chosen image_keys as create_video's catalog_image_keys.
get_reference_upload_page
ChatGPTLet the user upload reference photos (a person, a product) from their own device: returns a one-hour upload_page_url to give them. They pick up to 5 PNG/JPEG photos there and confirm they hold the rights to them. Use this in chat apps (Claude, ChatGPT), which can't upload files themselves. When the user says they're done, call check_reference_upload, then create_video(reference_upload_id=...).
get_reference_upload_url
ChatGPTFor clients that can make HTTP requests (e.g. Claude Code): a one-hour presigned upload for one reference image (png/jpg). Upload it with an HTTP multipart POST (every upload.fields entry as a form field, then the file as 'file'), then pass the returned reference_image_key in create_video's reference_image_keys. One call per image, up to 5. rights_confirmed must be true: first tell the user that they must own the image or hold the rights to use it and that Fabeo isn't responsible for infringement, and get their OK. Chat apps: use get_reference_upload_page or upload_reference_image.
get_scene_prompts
ChatGPTRead one scene as the renderer sees it: narration, visual description, reference images, and for each clip its MiniMax H3 positive_prompt (picture + the narration it speaks) and whether it opens on a user start frame. prompts_stale=true means a narration/visual edit is saved but not rendered yet (regenerate_scene mode='narration' applies it). Read this before writing a positive_prompt with edit_scene.
get_start_frame_upload_url
ChatGPTGet a one-hour upload URL for the user's own start frame (PNG) for one clip of a rendered scene. Upload it (multipart POST: upload.fields, then the file as 'file'), then call edit_scene(video_id, scene_id, clip_index, keyframe='upload', positive_prompt=...) and regenerate_scene with the mode it returns.
get_started
ChatGPTStart here with a new user, or when unsure what to make: what Fabeo makes, the ways to start (brief, template, product, own photos, own voice, quick clip), the seven video styles, limits, costs, template categories, and this account's plan and credits.
get_video
ChatGPTGet one video: status, playable video_url once rendered, and its scenes in order (scene_id, narration, visual description, duration, clip_count, reference images). Use get_scene_prompts for a rendered scene's clip prompts.
list_templates
ChatGPTBrowse Fabeo's proven template videos -- a quick way to a good video: copy one with create_from_template and adapt it to the user's product. Without video_type: the template categories and how many each has (e.g. product_demo, social_ad, ugc_style, before_after, animation, cinematic_trailer, explainer -- the result is authoritative). With video_type (optionally aspect_ratio '9:16' or '16:9'): each template's idea and a preview_url the user can watch -- show them as links and let the user choose.
list_videos
ChatGPTList videos in the account, newest first. Pass next_page_token from a previous response to fetch the next page. At most one of client_id, brand_id, product_id may be set.
regenerate_scene
ChatGPTRe-render some scenes of a FINISHED video and re-stitch it. Spends credits only for the re-rendered clip seconds of those scenes (~1 per second; key frames, voice and stitching are free; refunded in full if it fails) -- estimate_cost and the user's OK first. Use the regenerate_mode that edit_scene returned: - 'narration': applies narration / visual_description edits -- re-writes those scenes' prompts, then new voice, key frames and clips. - 'images': new key frames, then new clips (after keyframe='variant'). - 'clips': new clips from the current prompts, start frames and durations (after a positive_prompt, keyframe='upload' or duration_seconds edit); also a plain re-roll. Refuses 'images'/'clips' for a scene whose narration/visual edit is not applied yet. After set_scene_audio use 'narration'. If the user still dislikes the voice after a narration fix, suggest their own recording (get_audio_upload_page, then set_scene_audio) rather than more re-rolls.
regenerate_video
ChatGPTRe-render every clip of a finished video, or retry a video whose render failed at the clip/stitch step (e.g. it ran out of credits). Spends credits for the clip seconds it renders, charged like a new render: a failed clip render was already refunded in full, so the user does not pay twice. Scenes edited since the last render are re-written into their prompts first. For failures at other steps use retry_step.
retry_step
ChatGPTResume a FAILED video on the same job, keeping its script and every edit -- follow get_job_status's next_action. A failed script is written again for free (it then comes back for review). A render that failed before or during clip rendering or stitching carries on to the finished video and spends credits like the original render (estimate_cost and the user's OK first). After a content-policy block, rewrite the flagged scenes with edit_scene and pass content_edited=true. Not for a job that is still running or looks stuck.
set_scene_audio
ChatGPTReplace one scene's AI voice with the user's own recording -- the fallback when the AI voice mispronounces words, handles a regional language or accent badly, or adds gibberish or stray sounds that editing the narration didn't fix. The scene then shows the person speaking the recording, lip-synced, and is exactly as long as it (up to 60s). Pass audio_key or audio_upload_id (see get_audio_upload_page / upload_audio); narration (optional) is what the recording says, so the scene's text and visuals match it. remove=true goes back to the AI voice. Nothing is re-rendered here: on a rendered video apply it with regenerate_scene(mode='narration') (estimate_cost and the user's OK first).
upload_attached_image
ChatGPTFIRST CHOICE whenever the user attached, dropped or generated an image in this ChatGPT conversation: store it (PNG or JPEG, max 10 MB) by passing the attachment as image, and get a reference_image_key for create_video or edit_scene (start_frame_image_key, end_frame_image_key or reference_image_keys). Use this instead of get_reference_upload_page whenever the user attached the picture. rights_confirmed must be true: first tell the user By uploading, you confirm that you own these images or hold every right and permission needed to use them in videos (including the consent of any person shown), and that Fabeo is not responsible for any copyright, trademark, privacy or other infringement arising from images you upload. and get their OK.
upload_audio
ChatGPTStore the user's voice recording from a public https link to an MP3/WAV/M4A/OGG file (up to 60s, 25 MB); returns an audio_key for create_video_from_audio or set_scene_audio. rights_confirmed must be true: first tell the user By uploading, you confirm that you own this recording or hold every right and permission needed to use it in videos (including the consent of anyone whose voice is in it), and that Fabeo is not responsible for any copyright, privacy or other infringement arising from audio you upload. and get their OK. For a file on the user's device use get_audio_upload_page instead.
upload_reference_image
ChatGPTStore a reference image (a person's photo / character sheet, a product shot) from a public https link to a PNG or JPEG file (max 10 MB); returns a reference_image_key for create_video. rights_confirmed must be true: first tell the user By uploading, you confirm that you own these images or hold every right and permission needed to use them in videos (including the consent of any person shown), and that Fabeo is not responsible for any copyright, trademark, privacy or other infringement arising from images you upload. and get their OK. For photos on the user's device use get_reference_upload_page instead.