Skill instructions
ad-creatives1 KB
View saved version →
---
name: ad-creatives
description: "Four to six very different still ad directions for a product, service, venue, offer or event, one clean image each from ad archetypes, checked for invented claims and logo errors, then the picked ones finished as posters, still sets or ready-to-post images. Use when the user wants several ad ideas or concepts to choose from before committing; for one known poster with its copy, use static-ad, for product stills without concept exploration, product-photoshoot, and for an ad video, ugc-video."
---
# Ad creatives
Four to six very different still ad directions for a product, service, venue, offer or event, one clean image each from ad archetypes, checked for invented claims and logo errors, then the picked ones finished as posters, still sets or ready-to-post images.
Use this skill via the get_skill(skill_id="ad-creatives") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="ad-creatives", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
ad-remake847 Bytes
View saved version →
---
name: ad-remake
description: "An existing ad re-shot as new footage with the same beats, order, timing and pacing, and a new cast, product, language or market. Use when an ad that already works should be re-shot as all-new footage for another market, audience or language, keeping its beats and timing. To swap people or a product inside the original footage, use video-recast (one version) or ad-variations (several); to re-make it shot by shot with reference characters, cast-swap."
---
# Ad remake
An existing ad re-shot as new footage with the same beats, order, timing and pacing, and a new cast, product, language or market.
Use this skill via the get_skill(skill_id="ad-remake") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="ad-remake", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
ad-variations1.04 KB
View saved version →
---
name: ad-variations
description: "Several edited versions of one ad video that keep its motion, framing, cuts, timing and original soundtrack, each with its own swaps: new people, products, outfits, backgrounds or targeted on-screen text. Use when the user wants one or many test variations of one supplied video of 4 seconds or more, keeping its motion, framing, timing and audio while swapping people, products, outfits, backgrounds or a named piece of on-screen text: \"make 10 variations of my ad\", \"same ad with our new product\". For a single recast with mapped characters, use video-recast."
---
# Ad variations from one video
Several edited versions of one ad video that keep its motion, framing, cuts, timing and original soundtrack, each with its own swaps: new people, products, outfits, backgrounds or targeted on-screen text.
Use this skill via the get_skill(skill_id="ad-variations") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="ad-variations", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
ai-persona1 KB
View saved version →
---
name: ai-persona
description: "One consistent photoreal character (human, creature hybrid, caricature, a mix of other characters, or based on a consenting adult's photo) as a front/back/head character sheet, with its DNA card and a short persona picked from two options. Use when the user wants an AI influencer, virtual model or avatar, a creature, hybrid or caricature, a character mixed from others or made from a photo of themselves or a consenting adult, or a saved character remixed. Not for placing an existing character into scenes or videos."
---
# AI persona
One consistent photoreal character (human, creature hybrid, caricature, a mix of other characters, or based on a consenting adult's photo) as a front/back/head character sheet, with its DNA card and a short persona picked from two options.
Use this skill via the get_skill(skill_id="ai-persona") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="ai-persona", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
apply-design-feedback985 Bytes
View saved version →
---
name: apply-design-feedback
description: "A list of fixes for one ad, poster or thumbnail, from a design-feedback report, reviewers' notes or the user, made as targeted image edits, one fix at a time, each checked, with everything else kept as it was. Use when the user has feedback on an image and wants it acted on (\"fix these\", \"apply the comments\"); for the review itself use design-feedback, for the same edit across many photos batch-photo-edit, and for a new concept static-ad or thumbnail-set."
---
# Apply design feedback
A list of fixes for one ad, poster or thumbnail, from a design-feedback report, reviewers' notes or the user, made as targeted image edits, one fix at a time, each checked, with everything else kept as it was.
Use this skill via the get_skill(skill_id="apply-design-feedback") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="apply-design-feedback", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
aspect-reframe850 Bytes
View saved version →
---
name: aspect-reframe
description: "An existing video switched between 9:16 and 16:9 without cropping or shrinking it: the missing area is padded and painted in by a video model, piece by piece, with the original soundtrack kept. Use when the user wants a video they already have converted between 9:16 and 16:9 (landscape for YouTube, vertical for Reels or TikTok) while keeping the whole original picture."
---
# Reframe a video
An existing video switched between 9:16 and 16:9 without cropping or shrinking it: the missing area is padded and painted in by a video model, piece by piece, with the original soundtrack kept.
Use this skill via the get_skill(skill_id="aspect-reframe") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="aspect-reframe", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
audio-prompting690 Bytes
View saved version →
---
name: audio-prompting
description: "How to pick and keep a voice, direct delivery with inline cues, pin a language and accent, describe music, and find or make a sound effect. Use before generate_speech, generate_music or generate_sound_effect, and before choosing a voice with search_voices."
---
# Audio prompting guide
How to pick and keep a voice, direct delivery with inline cues, pin a language and accent, describe music, and find or make a sound effect.
Use this skill via the get_skill(skill_id="audio-prompting") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="audio-prompting", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
batch-photo-edit994 Bytes
View saved version →
---
name: batch-photo-edit
description: "The same edit or look across a set of the user's photos (warm and golden, moody, film, a cleaner background): a before/after preview on one photo first, then the rest with the identical prompt once the user confirms, each checked. Use when the user wants many photos edited the same way so the set looks cohesive; for retouching portraits use portrait-retouch, for several different fixes to one image apply-design-feedback, and for new product photos product-photoshoot."
---
# Batch photo edit
The same edit or look across a set of the user's photos (warm and golden, moody, film, a cleaner background): a before/after preview on one photo first, then the rest with the identical prompt once the user confirms, each checked.
Use this skill via the get_skill(skill_id="batch-photo-edit") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="batch-photo-edit", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
brand-kit1.11 KB
View saved version →
---
name: brand-kit
description: "A saved brand profile built from the user's website, brand name or a few answers: the brand name, colour palette in hex, fonts, tone, audience, products, logo, visual style, the claims it makes and what to do and avoid, kept in their project so later ads, photoshoots and videos reuse it without asking again. Use when the user wants Frameo to know their brand, gives a website, brand name or logo to work from, asks to set up, show or update their brand or see how other brands in their category look, or before work for a brand when no brand profile is saved yet."
---
# Brand kit
A saved brand profile built from the user's website, brand name or a few answers: the brand name, colour palette in hex, fonts, tone, audience, products, logo, visual style, the claims it makes and what to do and avoid, kept in their project so later ads, photoshoots and videos reuse it without asking again.
Use this skill via the get_skill(skill_id="brand-kit") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="brand-kit", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
brand-mockups1.01 KB
View saved version →
---
name: brand-mockups
description: "The user's own logo applied to photoreal mockups of merch, packaging, signage, screens and social images, every prompt held to one Brand Lock of exact colours, fonts and shape language, with a pilot mockup approved before the rest of the set. Use when the user has a logo or brand and wants to see it on products, merch, packaging, signs or social posts; not for designing a new logo, recolouring or vector files, an ad poster with copy (static-ad), or product stills without logo work (product-photoshoot)."
---
# Brand mockups
The user's own logo applied to photoreal mockups of merch, packaging, signage, screens and social images, every prompt held to one Brand Lock of exact colours, fonts and shape language, with a pilot mockup approved before the rest of the set.
Use this skill via the get_skill(skill_id="brand-mockups") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="brand-mockups", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
brand-story-explainer998 Bytes
View saved version →
---
name: brand-story-explainer
description: "A text-only prompt pack for a narrated 30–45 second flat-vector 2D ad in which a small character story sells the product: bibles with @tags, an 8-beat script, image prompts for the reference stills and 3–4 video blocks with one narrator voice. Use when the user wants the prompts for a 2D animated or flat-vector story ad in which characters act out the service under one narrator. It writes the prompts; it does not render the video."
---
# Brand story explainer
A text-only prompt pack for a narrated 30–45 second flat-vector 2D ad in which a small character story sells the product: bibles with @tags, an 8-beat script, image prompts for the reference stills and 3–4 video blocks with one narrator voice.
Use this skill via the get_skill(skill_id="brand-story-explainer") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="brand-story-explainer", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
bulk-from-table1.04 KB
View saved version →
---
name: bulk-from-table
description: "One image or video clip per row of a table the user pastes or opens in the chat (products, listings, names, offers): a template prompt with a slot per column, one pilot row checked, then the rest in waves of up to ten, with a row-by-row ledger. Use when the user has a list or spreadsheet and wants one asset per row: \"make an image for each product in this CSV\", \"one video per row\", \"generate these 30 listings\". Not for variants of one image (thumbnail-set, product-photoshoot) or variations of one ad video (ad-variations)."
---
# Bulk from a table
One image or video clip per row of a table the user pastes or opens in the chat (products, listings, names, offers): a template prompt with a slot per column, one pilot row checked, then the rest in waves of up to ten, with a row-by-row ledger.
Use this skill via the get_skill(skill_id="bulk-from-table") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="bulk-from-table", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
cast-swap1003 Bytes
View saved version →
---
name: cast-swap
description: "A source video re-made shot by shot with each mapped person replaced by your reference character: the same shots, story and dialogue, rebuilt from character-swapped keyframes animated chunk by chunk and joined. Use when the user has a video and wants it re-made with different people, each replaced by a reference character who performs and speaks the same dialogue in a new voice: \"replace the actors in this scene\", \"remix this video with my characters\". To keep the source's exact motion and soundtrack, use video-recast."
---
# Cast swap
A source video re-made shot by shot with each mapped person replaced by your reference character: the same shots, story and dialogue, rebuilt from character-swapped keyframes animated chunk by chunk and joined.
Use this skill via the get_skill(skill_id="cast-swap") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="cast-swap", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
character-sheet888 Bytes
View saved version →
---
name: character-sheet
description: "A consistent character across views and expressions, saved in the project so every later image and video can keep the same face. Use when the user wants a character, mascot, avatar or cast designed, or before multi-shot work with a recurring person or mascot, such as a series, a brand character, a story cast and similar. For an AI influencer or virtual model with a persona, use ai-persona; for every character, location and prop of a script made at once, story-assets."
---
# Character sheet
A consistent character across views and expressions, saved in the project so every later image and video can keep the same face.
Use this skill via the get_skill(skill_id="character-sheet") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="character-sheet", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
clay-explainer898 Bytes
View saved version →
---
name: clay-explainer
description: "A 70-second claymation explainer ad that teaches one topic or product through two contrasting clay characters: storyboarded shots stepped to stop-motion, one continuous narration, a music bed and burned-in stat cards. Use for a claymation, stop-motion or other handcrafted tabletop explainer or ad (plasticine, felt, toys) told through two contrasting characters, about 70 seconds with narration."
---
# Clay explainer
A 70-second claymation explainer ad that teaches one topic or product through two contrasting clay characters: storyboarded shots stepped to stop-motion, one continuous narration, a music bed and burned-in stat cards.
Use this skill via the get_skill(skill_id="clay-explainer") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="clay-explainer", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
continuity-check1.12 KB
View saved version →
---
name: continuity-check
description: "A lock map of what each reference must keep the same across shots (face, body, outfit, hair, product, place, camera, light, logo, type), the blocking unknowns, and a tested set of about three hard shots before the full run. Use before generating a series, a script with a cast, a product film or any many-shot job where a face, outfit, product, place or logo must not drift: \"keep her the same in every shot\", \"test it before you make all 20\". For shots already made, use review-shots; for one product that must stay exact, product-fidelity; to make the references, character-sheet or story-assets."
---
# Continuity check before a multi-shot job
A lock map of what each reference must keep the same across shots (face, body, outfit, hair, product, place, camera, light, logo, type), the blocking unknowns, and a tested set of about three hard shots before the full run.
Use this skill via the get_skill(skill_id="continuity-check") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="continuity-check", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
design-feedback944 Bytes
View saved version →
---
name: design-feedback
description: "A critique of the user's own ad, poster, thumbnail or video frame: ranked High/Med/Low findings, each located with a concrete fix, then what works, with an optional brand check of colours, fonts, logo and copy. Use when the user asks for feedback on, a review of or a brand check of a creative they made or brought; it changes nothing (apply-design-feedback makes the fixes), and shots generated for a plan are checked with review-shots instead."
---
# Design feedback
A critique of the user's own ad, poster, thumbnail or video frame: ranked High/Med/Low findings, each located with a concrete fix, then what works, with an optional brand check of colours, fonts, logo and copy.
Use this skill via the get_skill(skill_id="design-feedback") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="design-feedback", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
faceless-story-video905 Bytes
View saved version →
---
name: faceless-story-video
description: "A narrated explainer or story with no on-screen speaker: a locked look, reusable characters and places, multi-shot clips, a narrator voice, music and captions. Use for a narrated video with no one on screen: history and science shorts, explainers, kids' stories and songs, fairy tales, listicles, 'did you know' channels and similar. When the user brings a recorded voiceover or the finished narration of a story, use voiceover-to-video."
---
# Faceless story video
A narrated explainer or story with no on-screen speaker: a locked look, reusable characters and places, multi-shot clips, a narrator voice, music and captions.
Use this skill via the get_skill(skill_id="faceless-story-video") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="faceless-story-video", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
ffmpeg-recipes784 Bytes
View saved version →
---
name: ffmpeg-recipes
description: "Tested run_ffmpeg commands and the rules every job follows: trim, join, crossfade, fit a shape, mix and duck audio, burn text and subtitles, stills, encoding, speed, measuring a cut, and what each failure means. Use before writing a run_ffmpeg job, and again when one is refused or fails."
---
# ffmpeg recipes
Tested run_ffmpeg commands and the rules every job follows: trim, join, crossfade, fit a shape, mix and duck audio, burn text and subtitles, stills, encoding, speed, measuring a cut, and what each failure means.
Use this skill via the get_skill(skill_id="ffmpeg-recipes") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="ffmpeg-recipes", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
get-started663 Bytes
View saved version →
---
name: get-started
description: "Start here: learn what can be made, pick a first idea and take it from a brief to a finished video or image. Use right after this app is connected, when someone asks what it can do or how to begin, and before the first generation of a new piece of work, to agree its brief and plan."
---
# Get started
Start here: learn what can be made, pick a first idea and take it from a brief to a finished video or image.
Use this skill via the get_skill(skill_id="get-started") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="get-started", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
highlight-reel965 Bytes
View saved version →
---
name: highlight-reel
description: "A long video the user already has cut into one highlight reel or several stand-alone shorts: the strongest moments picked from its transcript and shots, cropped around the speaker for 9:16, joined, with optional captions and a hook. Use when the user wants the best parts of an existing long video (interview, podcast, talk, vlog, event) chosen for them, as a reel or as shorts; when the user gives the cut points themselves, use subtitles-and-cut."
---
# Highlight reel
A long video the user already has cut into one highlight reel or several stand-alone shorts: the strongest moments picked from its transcript and shots, cropped around the speaker for 9:16, joined, with optional captions and a hook.
Use this skill via the get_skill(skill_id="highlight-reel") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="highlight-reel", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
hindi-redub1003 Bytes
View saved version →
---
name: hindi-redub
description: "The user's videos with their Hindi or Hinglish dialogue re-voiced from the script: every line regenerated in Devanagari with the right emotion, laid in frame-exact, on-camera shots lip-synced, each video kept to its exact length. Use when the user's videos have Hindi or Hinglish dialogue that is mispronounced or sounds flat and should be redone from their script: \"fix the hindi dialogue\", \"redub hindi\". Not for translating a video into Hindi (video-dub), new footage or other languages."
---
# Fix Hindi dialogue
The user's videos with their Hindi or Hinglish dialogue re-voiced from the script: every line regenerated in Devanagari with the right emotion, laid in frame-exact, on-camera shots lip-synced, each video kept to its exact length.
Use this skill via the get_skill(skill_id="hindi-redub") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="hindi-redub", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
image-prompting869 Bytes
View saved version →
---
name: image-prompting
description: "How to write an image prompt or an edit: which image is the input and which the reference, what to keep, camera angles, shape, model and resolution, plus tested prompts for a new camera angle, relighting, reframing and a storyboard sketch. Use before writing a prompt for generate_images, and when an image or an edit comes back wrong."
---
# Image prompting guide
How to write an image prompt or an edit: which image is the input and which the reference, what to keep, camera angles, shape, model and resolution, plus tested prompts for a new camera angle, relighting, reframing and a storyboard sketch.
Use this skill via the get_skill(skill_id="image-prompting") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="image-prompting", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
location-enhance1011 Bytes
View saved version →
---
name: location-enhance
description: "A location image, or every panel of a location sheet, rebuilt as a natural, real-scale photograph of the same place: every element kept in position, its proportions, materials and light made real, one 2K image per panel. Use when a location image or sheet looks fake, AI-textured or stylized (a clay set, a miniature, a cartoon or low-poly scene) and should become a real place: \"make this place look real\", \"better quality location\". Not for characters, or a small edit to one element."
---
# Enhance a location image
A location image, or every panel of a location sheet, rebuilt as a natural, real-scale photograph of the same place: every element kept in position, its proportions, materials and light made real, one 2K image per panel.
Use this skill via the get_skill(skill_id="location-enhance") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="location-enhance", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
multi-cam-edit913 Bytes
View saved version →
---
name: multi-cam-edit
description: "One single-angle video turned into a multi-camera version: the same people, actions, dialogue and setting shown from several camera angles with hard cuts, the original soundtrack kept. Use when the user has a video shot from one angle and wants it to look like a multi-camera shoot: \"add more camera angles\", \"make it multicam\", for a talking head, interview, performance or product clip. Not for footage that already has several angles."
---
# Multi-camera edit
One single-angle video turned into a multi-camera version: the same people, actions, dialogue and setting shown from several camera angles with hard cuts, the original soundtrack kept.
Use this skill via the get_skill(skill_id="multi-cam-edit") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="multi-cam-edit", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
portrait-retouch1.14 KB
View saved version →
---
name: portrait-retouch
description: "A portrait photo the user already has, retouched naturally: light, colour and a named blemish fixed, an optional look and background softening, with the face, skin texture and marks kept as they are, one photo or a whole shoot under one look. Use when the user wants their own or a consenting person's portrait or headshot photo polished without changing who they are: \"retouch my photo\", \"fix the lighting on this portrait\", \"clean up these headshots\". Not for changing age, body, hair or expression (a plain edit with image-prompting), a new character from a photo (ai-persona) or a fake-looking place (location-enhance)."
---
# Portrait retouch
A portrait photo the user already has, retouched naturally: light, colour and a named blemish fixed, an optional look and background softening, with the face, skin texture and marks kept as they are, one photo or a whole shoot under one look.
Use this skill via the get_skill(skill_id="portrait-retouch") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="portrait-retouch", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
presenter-overlay945 Bytes
View saved version →
---
name: presenter-overlay
description: "Put the user (from their photo) or a designed presenter into a video they already have, as a picture-in-picture narrator speaking the script: cut out in a corner, in a round badge, or full height, across the whole video. Use for 'put me in this video', a talking head over the user's own clip, or a narrator added on screen to existing footage; for a presenter alone with no footage of the user's, use talking-presenter."
---
# Presenter over your video
Put the user (from their photo) or a designed presenter into a video they already have, as a picture-in-picture narrator speaking the script: cut out in a corner, in a round badge, or full height, across the whole video.
Use this skill via the get_skill(skill_id="presenter-overlay") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="presenter-overlay", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
product-fidelity1.04 KB
View saved version →
---
name: product-fidelity
description: "A checkable fact sheet of a product (shape, colour, packaging, label, logo, material) with the views its photos support, which parts of each shot must be the real image or file and which can be generated, and a three-shot test before the set or film. Use before making several images or clips of one product that must stay exactly right: \"the label has to be perfect in every shot\", \"keep our bottle exact in the video\". It plans and tests; the shots come from product-photoshoot, ugc-video or static-ad."
---
# Product fidelity across shots
A checkable fact sheet of a product (shape, colour, packaging, label, logo, material) with the views its photos support, which parts of each shot must be the real image or file and which can be generated, and a three-shot test before the set or film.
Use this skill via the get_skill(skill_id="product-fidelity") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="product-fidelity", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
product-photoshoot764 Bytes
View saved version →
---
name: product-photoshoot
description: "Product stills from a few photos or a description: packshots, lifestyle scenes, close-ups with hands, hero banners, carousels, ad packs, on-model shots, conceptual stills and restyles. Use for e-commerce listings and ad stills; for video, use `ugc-video` or a camera-move preset."
---
# Product photoshoot
Product stills from a few photos or a description: packshots, lifestyle scenes, close-ups with hands, hero banners, carousels, ad packs, on-model shots, conceptual stills and restyles.
Use this skill via the get_skill(skill_id="product-photoshoot") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="product-photoshoot", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
reference-pacing944 Bytes
View saved version →
---
name: reference-pacing
description: "A reference reel's pacing (cuts, flashes, shake bursts, camera moves and hits per second) read into a timing map and added to the user's edit in post: punch-ins, shake, flashes and speed ramps on clean in-points. Use when an edit or its shots exist and should hit with a reference reel's energy because the generated clips came out calmer; to make the shots from a reference use `reference-remix`, for music and sound effects `sound-pass`."
---
# Reference pacing
A reference reel's pacing (cuts, flashes, shake bursts, camera moves and hits per second) read into a timing map and added to the user's edit in post: punch-ins, shake, flashes and speed ramps on clean in-points.
Use this skill via the get_skill(skill_id="reference-pacing") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="reference-pacing", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
reference-remix705 Bytes
View saved version →
---
name: reference-remix
description: "A new edit of the user's own subject in the visual language of a reference video: its pacing, shot sizes, camera moves, transitions, text style and colour. Use when the user points at an edit they like and wants that feel for their own product, person or story."
---
# Reference remix
A new edit of the user's own subject in the visual language of a reference video: its pacing, shot sizes, camera moves, transitions, text style and colour.
Use this skill via the get_skill(skill_id="reference-remix") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="reference-remix", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
resize-for-social1.09 KB
View saved version →
---
name: resize-for-social
description: "One image or video made ready for every social platform at its exact size (Instagram, TikTok, LinkedIn, Facebook, X, YouTube, Snapchat, Pinterest, Threads), cropped around the face or product, with an image extended by an image model where a size needs more picture, and named per platform. Use when the user wants an existing image or video in one or more platform sizes or at an exact size or shape; for a video switched between 9:16 and 16:9 with the new area painted in, use aspect-reframe, and for a new thumbnail, thumbnail-set."
---
# Resize for social
One image or video made ready for every social platform at its exact size (Instagram, TikTok, LinkedIn, Facebook, X, YouTube, Snapchat, Pinterest, Threads), cropped around the face or product, with an image extended by an image model where a size needs more picture, and named per platform.
Use this skill via the get_skill(skill_id="resize-for-social") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="resize-for-social", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
review-shots857 Bytes
View saved version →
---
name: review-shots
description: "Check every image and clip against the plan before building on it — the right people, the asked-for action, clean hands and text, lips in sync — and retake a failed shot once. Use on shots made earlier when the user asks for a check; the video and photoshoot skills include the same check."
---
# Review shots
Check every image and clip against the plan before building on it — the right people, the asked-for action, clean hands and text, lips in sync — and retake a failed shot once.
Use this skill via the get_skill(skill_id="review-shots") tool call of this app: call it first; it returns the written steps of this check, as plain text. Its reference pages, also plain text, come from get_skill(skill_id="review-shots", path=...) when the steps name them. Nothing in this skill is code to install or run.
Referenced files: 2 — file references are metadata, not package contents.
script-to-video833 Bytes
View saved version →
---
name: script-to-video
description: "Turn a script into a finished video, such as a micro-drama — a consistent cast, one shot per beat, characters speaking their lines, cut together with subtitles. Use when a script should become a video with speaking characters: short films, series, micro-dramas, dialogue ads, sketches, kids' stories and similar. With only an idea or a story, video-script writes the script first."
---
# Script to video
Turn a script into a finished video, such as a micro-drama — a consistent cast, one shot per beat, characters speaking their lines, cut together with subtitles.
Use this skill via the get_skill(skill_id="script-to-video") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="script-to-video", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
shot-plan1.02 KB
View saved version →
---
name: shot-plan
description: "A camera plan for one scene before it is animated: a shot table with length, size, angle, camera position, lens, one move and light per shot, the reference views it needs, and one separate opening frame per shot with its video prompt line. Use when the user wants a scene's shots, camera positions and moves planned and seen as frames before any video is made: \"plan the shots\", \"camera plan for this scene\". For one grid image of panels, use storyboard-sheet; for a whole script made into a finished video, script-to-video."
---
# Shot plan for a scene
A camera plan for one scene before it is animated: a shot table with length, size, angle, camera position, lens, one move and light per shot, the reference views it needs, and one separate opening frame per shot with its video prompt line.
Use this skill via the get_skill(skill_id="shot-plan") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="shot-plan", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
sound-pass815 Bytes
View saved version →
---
name: sound-pass
description: "A finished video scored and sweetened: one continuous music bed, a few well-placed library sound effects and one mix ducked under the dialogue, checked by a listen before delivery. Use when a finished video needs its whole sound pass, music and sound effects mixed under its dialogue; for trims, captions or a plain music bed on raw footage, use `subtitles-and-cut`."
---
# Sound pass
A finished video scored and sweetened: one continuous music bed, a few well-placed library sound effects and one mix ducked under the dialogue, checked by a listen before delivery.
Use this skill via the get_skill(skill_id="sound-pass") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="sound-pass", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
static-ad914 Bytes
View saved version →
---
name: static-ad
description: "An on-brand static ad poster with its headline and CTA rendered in the image, recreating a reference ad's layout or built from scratch, from the brand's guidelines or web research, with optional motion added afterwards. Use when the user wants an ad poster or banner with copy for their brand, or a reference ad remade with their product; for product stills without ad copy, use product-photoshoot, and for an ad video, ugc-video."
---
# Static ad
An on-brand static ad poster with its headline and CTA rendered in the image, recreating a reference ad's layout or built from scratch, from the brand's guidelines or web research, with optional motion added afterwards.
Use this skill via the get_skill(skill_id="static-ad") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="static-ad", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
story-assets1 KB
View saved version →
---
name: story-assets
description: "A script's whole visual library from one approved roster: identity references and an outfit datasheet for every character, three views and a datasheet for every location, and the story's props, saved to the project and its cast. Use when the user has a script or story and wants every character, location and prop it needs made as reference images in one go before shooting: \"create all the characters and locations for my script\". For a single character, use character-sheet; for the shots and the video, script-to-video."
---
# Story assets
A script's whole visual library from one approved roster: identity references and an outfit datasheet for every character, three views and a datasheet for every location, and the story's props, saved to the project and its cast.
Use this skill via the get_skill(skill_id="story-assets") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="story-assets", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
storyboard-sheet794 Bytes
View saved version →
---
name: storyboard-sheet
description: "One storyboard contact-sheet image of a scene or short sequence: 9, 12 or 16 numbered panels, each a planned and directed frame, drawn in one consistent art style. Use when the user asks for a grid storyboard or storyboard sheet of one scene or a short sequence, as a single image of numbered panels rather than separate shots or a video."
---
# Storyboard sheet
One storyboard contact-sheet image of a scene or short sequence: 9, 12 or 16 numbered panels, each a planned and directed frame, drawn in one consistent art style.
Use this skill via the get_skill(skill_id="storyboard-sheet") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="storyboard-sheet", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
subtitles-and-cut737 Bytes
View saved version →
---
name: subtitles-and-cut
description: "Edit footage the user already has: trim, resize to 9:16 or 16:9, join clips, add a music bed and add captions from the user's timings or text, or from a transcription of the footage. Use when the footage already exists and only needs editing, not new scenes."
---
# Subtitles and cut
Edit footage the user already has: trim, resize to 9:16 or 16:9, join clips, add a music bed and add captions from the user's timings or text, or from a transcription of the footage.
Use this skill via the get_skill(skill_id="subtitles-and-cut") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="subtitles-and-cut", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
talking-head-overlays1.03 KB
View saved version →
---
name: talking-head-overlays
description: "The user's vertical talking-head clip finished with effects timed to what is said: styled captions, freeze-frames with name tags, number cards, an early before/after grade wipe, punch-ins and shakes on hits, and library sound effects on every moment. Use when the user has a clip of one person talking to camera and wants reel-style edit effects on it; for plain trims, joins or simple captions use `subtitles-and-cut`, and for music and sound effects alone on a finished film use `sound-pass`."
---
# Talking-head overlays
The user's vertical talking-head clip finished with effects timed to what is said: styled captions, freeze-frames with name tags, number cards, an early before/after grade wipe, punch-ins and shakes on hits, and library sound effects on every moment.
Use this skill via the get_skill(skill_id="talking-head-overlays") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="talking-head-overlays", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
talking-presenter791 Bytes
View saved version →
---
name: talking-presenter
description: "One person on camera delivers a script: a face (photo or designed), a fixed voice, lipsynced delivery, optional title card, cut to length. Use for explainers, founder messages, course intros and announcements delivered by one person on camera. For a presenter over footage the user already has, use presenter-overlay; for two hosts on a set, video-podcast."
---
# Talking presenter
One person on camera delivers a script: a face (photo or designed), a fixed voice, lipsynced delivery, optional title card, cut to length.
Use this skill via the get_skill(skill_id="talking-presenter") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="talking-presenter", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
thumbnail-set691 Bytes
View saved version →
---
name: thumbnail-set
description: "Three to five thumbnail concepts for a video, then variants of the chosen one: a face or product, a short title treatment, high contrast, delivered at upload size. Use for YouTube thumbnails, video covers, Reels or Shorts covers and A/B test sets."
---
# Thumbnail set
Three to five thumbnail concepts for a video, then variants of the chosen one: a face or product, a short title treatment, high contrast, delivered at upload size.
Use this skill via the get_skill(skill_id="thumbnail-set") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="thumbnail-set", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
translate-on-screen-text1009 Bytes
View saved version →
---
name: translate-on-screen-text
description: "The words of an existing static design (ad, poster, post, banner) regenerated in another language with the picture and layout kept, or a video given burned-in subtitles in another language with its sound untouched, from an approved source-and-target sheet. Use when the user wants a design's text or a video's subtitles in another language without any voice work; to re-voice a video's speech in another language, use video-dub."
---
# Translate on-screen text
The words of an existing static design (ad, poster, post, banner) regenerated in another language with the picture and layout kept, or a video given burned-in subtitles in another language with its sound untouched, from an approved source-and-target sheet.
Use this skill via the get_skill(skill_id="translate-on-screen-text") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="translate-on-screen-text", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
ugc-product-video7.61 KB
View saved version →
---
name: ugc-product-video
description: "Product-only UGC: handheld-style shots of the product with a voiceover, no creator on camera."
---
Needs: product photos; the one claim
Credits: about 390–470 for 30 seconds, plus a retake allowance of 15% of the image and clip total
Time: 10–20 minutes, plus about a minute per clip for the review
## When
When there is no creator to show, or the brand wants the product alone. Not for a person on
camera (`ugc-review-video`).
## Ask first
1. **The product** — photos (`show_upload` for files on the user's device, `import_media_url` for a web link; images attached to the chat do not reach Frameo, and `create_upload_url` is for clients that send the file themselves), the name, and the one thing the video
must say about it.
2. **The voice** — the tone of the voiceover (warm, brisk, playful) and its language.
3. **Frame and length** — 9:16, 15 / 30 / 45 s.
Say the rough cost in the same turn as the questions; the user's yes comes with the plan (Steps, **The plan first.**).
## Steps
**Before any paid step.** The project: `list_projects` (or `create_project`) gives the `project_id`
and, when the project has several modules, the `module_id`; every call below that takes a project —
`estimate_cost` included — gets that same pair; without it those tools answer `project_needed`. The
quote: one `estimate_cost(items=[…])` prices a stage in one call, with the same project, model,
size and number of `image_urls` or `reference_image_urls` (`reference_count`) as each generate call, and returns a `quote_id` per item plus the total. Each generate call
then passes its own item's `quote_id` and `confirmed_by_user=true`. A quote is single-use and lasts
15 minutes, so a long plan is priced stage by stage, right before each stage runs; a stage that
comes to more than the user approved is asked about again first. Generate calls return
`generation_ids`; `wait_task` returns the links, and `show_generations` shows each stage's running
and finished work in one card where the chat app displays Frameo cards: all the stage's ids at
once, before its first `wait_task`, and the finished result with `final: true`.
**The plan first.** Before the first paid call, the plan goes to the user in the chat as plain
text: what will be made, in order, one line per generation (for a script, the shot list; for a
set, each shot), with each line's credits and the total. The credits come from
`estimate_cost(items=[…])`, up to 10 items a call, so a long plan takes several calls; those
quotes may expire unused, since each stage is quoted again right before it runs. Nothing is
generated until the user says yes. The user can drop or change lines; a changed line is priced
again.
**Canvas rows.** Pass `shot_number` on every `generate_image` and `generate_video` of a shot (1, 2, 3… in story order; the same number for a retake and for that shot's video), so each shot gets its own row on the Frameo canvas. Cast, prop and location references take `placement_kind` (`character`, `prop` or `location`) and the subject's name as `placement_group` instead: they sit on their own board, and the shots built from them do not pile into their row.
**Review (on by default).** Every still and clip is checked before it is built on, and the
finished cut before it is handed over, following `review-shots`: its checklist and steps come
from `get_skill("review-shots")`, loaded before step 1. The checks run on the step-1 still
before step 2 builds on it, after step 2 (`review-shots` steps 0, 1 and 4), after step 4 (steps 2 and 4) and after the cut (step 5); the
product's label, shape and colour are checked against the product photos in every shot, and a
failed shot is retaken once. No clip has speech on screen, so every check is free; a retake is
paid like the shot it replaces, from the retake allowance the plan carries as its own line for
the user to approve. The user can turn the review off.
**0. Script.** Write a voiceover script of the length chosen Preview of the first 4,000 characters. Read full instructions →
Referenced files: 2 — file references are metadata, not package contents.
ugc-review-video9.29 KB
View saved version →
---
name: ugc-review-video
description: "The default UGC ad, a creator on camera reviewing a product in phone-video style: hook, one honest claim, product in hand, a close — lipsynced delivery."
---
Needs: product photos; the one claim; a creator photo or a description
Credits: about 500–650 for 30 seconds, plus a retake allowance of 15% of the image and clip total
Time: 15–25 minutes, plus about a minute per clip for the review
## When
The default UGC ask: a person talking about a product as if to their followers. Not for
product-only shots (`ugc-product-video`), unboxing (`ugc-unboxing-video`), wearing it
(`ugc-try-on-video`) or a how-to (`ugc-tutorial-video`).
## Ask first
1. **The product** — photos (`show_upload` for files on the user's device, `import_media_url` for a web link; images attached to the chat do not reach Frameo, and `create_upload_url` is for clients that send the file themselves), the name, and the one thing the video
must say about it.
2. **The creator** — design one (age, vibe, setting), use the user's photo, or reuse a creator
already in the project (`list_characters`: its first image is the step-1 still and its
`voice_id` the voice in step 3, with nothing to generate or choose).
3. **Frame and length** — 9:16, 15 / 30 / 45 s.
Say the rough cost in the same turn as the questions; the user's yes comes with the plan (Steps, **The plan first.**).
## Steps
**Before any paid step.** The project: `list_projects` (or `create_project`) gives the `project_id`
and, when the project has several modules, the `module_id`; every call below that takes a project —
`estimate_cost` included — gets that same pair; without it those tools answer `project_needed`. The
quote: one `estimate_cost(items=[…])` prices a stage in one call, with the same project, model,
size and number of `image_urls` or `reference_image_urls` (`reference_count`) as each generate call, and returns a `quote_id` per item plus the total. Each generate call
then passes its own item's `quote_id` and `confirmed_by_user=true`. A quote is single-use and lasts
15 minutes, so a long plan is priced stage by stage, right before each stage runs; a stage that
comes to more than the user approved is asked about again first. Generate calls return
`generation_ids`; `wait_task` returns the links, and `show_generations` shows each stage's running
and finished work in one card where the chat app displays Frameo cards: all the stage's ids at
once, before its first `wait_task`, and the finished result with `final: true`.
**The plan first.** Before the first paid call, the plan goes to the user in the chat as plain
text: what will be made, in order, one line per generation (for a script, the shot list; for a
set, each shot), with each line's credits and the total. The credits come from
`estimate_cost(items=[…])`, up to 10 items a call, so a long plan takes several calls; those
quotes may expire unused, since each stage is quoted again right before it runs. Nothing is
generated until the user says yes. The user can drop or change lines; a changed line is priced
again.
**Canvas rows.** Pass `shot_number` on every `generate_image` and `generate_video` of a shot (1, 2, 3… in story order; the same number for a retake and for that shot's video), so each shot gets its own row on the Frameo canvas. Cast, prop and location references take `placement_kind` (`character`, `prop` or `location`) and the subject's name as `placement_group` instead: they sit on their own board, and the shots built from them do not pile into their row.
**Review (on by default).** Every still and clip is checked before it is built on, and the
finished cut before it is handed over, following `review-shots`: its checklist and steps come
from `get_skill("review-shots")`, loaded before step 1. The checks run on the step-1 still
before step 2 builds on it, after step 2 (`review-shots` steps 0, 1 and 4), after step 4 (steps 2 and 4) and after the cut (step 5); the
product's label, shape and cPreview of the first 4,000 characters. Read full instructions →
Referenced files: 2 — file references are metadata, not package contents.
ugc-try-on-video9.3 KB
View saved version →
---
name: ugc-try-on-video
description: "Wearing and fit-check for clothing, jewellery, glasses or shoes: the creator in the item, a turn, a close-up, a verdict."
---
Needs: product photos; a creator photo (the person who wears it) or a description; size or fit notes
Credits: about 500–650 for 30 seconds, plus a retake allowance of 15% of the image and clip total
Time: 15–25 minutes, plus about a minute per clip for the review
## When
Wearable products. The creator's photo matters more here than anywhere else: the item has
to sit on a consistent person. Not for products that are held or used (`ugc-review-video`).
## Ask first
1. **The product** — photos (`show_upload` for files on the user's device, `import_media_url` for a web link; images attached to the chat do not reach Frameo, and `create_upload_url` is for clients that send the file themselves), the name, and the one thing the video
must say about it.
2. **The creator** — design one (age, vibe, setting), use the user's photo, or reuse a creator
already in the project (`list_characters`: its first image is the step-1 still and its
`voice_id` the voice in step 3, with nothing to generate or choose).
3. **Frame and length** — 9:16, 15 / 30 / 45 s.
Say the rough cost in the same turn as the questions; the user's yes comes with the plan (Steps, **The plan first.**).
## Steps
**Before any paid step.** The project: `list_projects` (or `create_project`) gives the `project_id`
and, when the project has several modules, the `module_id`; every call below that takes a project —
`estimate_cost` included — gets that same pair; without it those tools answer `project_needed`. The
quote: one `estimate_cost(items=[…])` prices a stage in one call, with the same project, model,
size and number of `image_urls` or `reference_image_urls` (`reference_count`) as each generate call, and returns a `quote_id` per item plus the total. Each generate call
then passes its own item's `quote_id` and `confirmed_by_user=true`. A quote is single-use and lasts
15 minutes, so a long plan is priced stage by stage, right before each stage runs; a stage that
comes to more than the user approved is asked about again first. Generate calls return
`generation_ids`; `wait_task` returns the links, and `show_generations` shows each stage's running
and finished work in one card where the chat app displays Frameo cards: all the stage's ids at
once, before its first `wait_task`, and the finished result with `final: true`.
**The plan first.** Before the first paid call, the plan goes to the user in the chat as plain
text: what will be made, in order, one line per generation (for a script, the shot list; for a
set, each shot), with each line's credits and the total. The credits come from
`estimate_cost(items=[…])`, up to 10 items a call, so a long plan takes several calls; those
quotes may expire unused, since each stage is quoted again right before it runs. Nothing is
generated until the user says yes. The user can drop or change lines; a changed line is priced
again.
**Canvas rows.** Pass `shot_number` on every `generate_image` and `generate_video` of a shot (1, 2, 3… in story order; the same number for a retake and for that shot's video), so each shot gets its own row on the Frameo canvas. Cast, prop and location references take `placement_kind` (`character`, `prop` or `location`) and the subject's name as `placement_group` instead: they sit on their own board, and the shots built from them do not pile into their row.
**Review (on by default).** Every still and clip is checked before it is built on, and the
finished cut before it is handed over, following `review-shots`: its checklist and steps come
from `get_skill("review-shots")`, loaded before step 1. The checks run on the step-1 still
before step 2 builds on it, after step 2 (`review-shots` steps 0, 1 and 4), after step 4 (steps 2 and 4) and after the cut (step 5); the
product's label, shape and colour are checked against the product photos in evePreview of the first 4,000 characters. Read full instructions →
Referenced files: 2 — file references are metadata, not package contents.
ugc-tutorial-video9.35 KB
View saved version →
---
name: ugc-tutorial-video
description: "A step-by-step how-to with \"Step N\" captions: setup, each step shown, the result."
---
Needs: product photos; the steps in order; a creator photo or a description
Credits: about 600–750 for 45 seconds, plus a retake allowance of 15% of the image and clip total
Time: 15–30 minutes, plus about a minute per clip for the review
## When
When the product needs showing how: setup, a routine, a recipe, an app flow. Not for a
review (`ugc-review-video`).
## Ask first
1. **The product** — photos (`show_upload` for files on the user's device, `import_media_url` for a web link; images attached to the chat do not reach Frameo, and `create_upload_url` is for clients that send the file themselves), the name, and the one thing the video
must say about it.
2. **The creator** — design one (age, vibe, setting), use the user's photo, or reuse a creator
already in the project (`list_characters`: its first image is the step-1 still and its
`voice_id` the voice in step 3, with nothing to generate or choose).
3. **Frame and length** — 9:16, 15 / 30 / 45 s.
Say the rough cost in the same turn as the questions; the user's yes comes with the plan (Steps, **The plan first.**).
## Steps
**Before any paid step.** The project: `list_projects` (or `create_project`) gives the `project_id`
and, when the project has several modules, the `module_id`; every call below that takes a project —
`estimate_cost` included — gets that same pair; without it those tools answer `project_needed`. The
quote: one `estimate_cost(items=[…])` prices a stage in one call, with the same project, model,
size and number of `image_urls` or `reference_image_urls` (`reference_count`) as each generate call, and returns a `quote_id` per item plus the total. Each generate call
then passes its own item's `quote_id` and `confirmed_by_user=true`. A quote is single-use and lasts
15 minutes, so a long plan is priced stage by stage, right before each stage runs; a stage that
comes to more than the user approved is asked about again first. Generate calls return
`generation_ids`; `wait_task` returns the links, and `show_generations` shows each stage's running
and finished work in one card where the chat app displays Frameo cards: all the stage's ids at
once, before its first `wait_task`, and the finished result with `final: true`.
**The plan first.** Before the first paid call, the plan goes to the user in the chat as plain
text: what will be made, in order, one line per generation (for a script, the shot list; for a
set, each shot), with each line's credits and the total. The credits come from
`estimate_cost(items=[…])`, up to 10 items a call, so a long plan takes several calls; those
quotes may expire unused, since each stage is quoted again right before it runs. Nothing is
generated until the user says yes. The user can drop or change lines; a changed line is priced
again.
**Canvas rows.** Pass `shot_number` on every `generate_image` and `generate_video` of a shot (1, 2, 3… in story order; the same number for a retake and for that shot's video), so each shot gets its own row on the Frameo canvas. Cast, prop and location references take `placement_kind` (`character`, `prop` or `location`) and the subject's name as `placement_group` instead: they sit on their own board, and the shots built from them do not pile into their row.
**Review (on by default).** Every still and clip is checked before it is built on, and the
finished cut before it is handed over, following `review-shots`: its checklist and steps come
from `get_skill("review-shots")`, loaded before step 1. The checks run on the step-1 still
before step 2 builds on it, after step 2 (`review-shots` steps 0, 1 and 4), after step 4 (steps 2 and 4) and after the cut (step 5); the
product's label, shape and colour are checked against the product photos in every shot, and a
failed shot is retaken once. The plan carries the retake allowance as its own line for the user
to approve withPreview of the first 4,000 characters. Read full instructions →
Referenced files: 2 — file references are metadata, not package contents.
ugc-unboxing-video9.26 KB
View saved version →
---
name: ugc-unboxing-video
description: "Unboxing / first-reaction / haul: the creator opens the package, reacts, shows what is inside."
---
Needs: product and packaging photos; a creator photo or a description
Credits: about 500–650 for 30 seconds, plus a retake allowance of 15% of the image and clip total
Time: 15–25 minutes, plus about a minute per clip for the review
## When
The package matters: first impressions, hauls, subscription boxes. Not for a plain
review (`ugc-review-video`).
## Ask first
1. **The product** — photos (`show_upload` for files on the user's device, `import_media_url` for a web link; images attached to the chat do not reach Frameo, and `create_upload_url` is for clients that send the file themselves), the name, and the one thing the video
must say about it.
2. **The creator** — design one (age, vibe, setting), use the user's photo, or reuse a creator
already in the project (`list_characters`: its first image is the step-1 still and its
`voice_id` the voice in step 3, with nothing to generate or choose).
3. **Frame and length** — 9:16, 15 / 30 / 45 s.
Say the rough cost in the same turn as the questions; the user's yes comes with the plan (Steps, **The plan first.**).
## Steps
**Before any paid step.** The project: `list_projects` (or `create_project`) gives the `project_id`
and, when the project has several modules, the `module_id`; every call below that takes a project —
`estimate_cost` included — gets that same pair; without it those tools answer `project_needed`. The
quote: one `estimate_cost(items=[…])` prices a stage in one call, with the same project, model,
size and number of `image_urls` or `reference_image_urls` (`reference_count`) as each generate call, and returns a `quote_id` per item plus the total. Each generate call
then passes its own item's `quote_id` and `confirmed_by_user=true`. A quote is single-use and lasts
15 minutes, so a long plan is priced stage by stage, right before each stage runs; a stage that
comes to more than the user approved is asked about again first. Generate calls return
`generation_ids`; `wait_task` returns the links, and `show_generations` shows each stage's running
and finished work in one card where the chat app displays Frameo cards: all the stage's ids at
once, before its first `wait_task`, and the finished result with `final: true`.
**The plan first.** Before the first paid call, the plan goes to the user in the chat as plain
text: what will be made, in order, one line per generation (for a script, the shot list; for a
set, each shot), with each line's credits and the total. The credits come from
`estimate_cost(items=[…])`, up to 10 items a call, so a long plan takes several calls; those
quotes may expire unused, since each stage is quoted again right before it runs. Nothing is
generated until the user says yes. The user can drop or change lines; a changed line is priced
again.
**Canvas rows.** Pass `shot_number` on every `generate_image` and `generate_video` of a shot (1, 2, 3… in story order; the same number for a retake and for that shot's video), so each shot gets its own row on the Frameo canvas. Cast, prop and location references take `placement_kind` (`character`, `prop` or `location`) and the subject's name as `placement_group` instead: they sit on their own board, and the shots built from them do not pile into their row.
**Review (on by default).** Every still and clip is checked before it is built on, and the
finished cut before it is handed over, following `review-shots`: its checklist and steps come
from `get_skill("review-shots")`, loaded before step 1. The checks run on the step-1 still
before step 2 builds on it, after step 2 (`review-shots` steps 0, 1 and 4), after step 4 (steps 2 and 4) and after the cut (step 5); the
product's label, shape and colour are checked against the product photos in every shot, and a
failed shot is retaken once. The plan carries the retake allowance as its own line for the user
to approve witPreview of the first 4,000 characters. Read full instructions →
Referenced files: 2 — file references are metadata, not package contents.
ugc-video720 Bytes
View saved version →
---
name: ugc-video
description: "A phone-video style product ad in one of five formats: a creator review, the product alone with a voiceover, an unboxing, a try-on or a step-by-step tutorial. Use for UGC and creator-style ads, from a quick review to a how-to, or a week of posts from one creator; pick the format from the request."
---
# UGC video
A phone-video style product ad in one of five formats: a creator review, the product alone with a voiceover, an unboxing, a try-on or a step-by-step tutorial.
Use this skill via the get_skill(skill_id="ugc-video") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="ugc-video", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
video-dub931 Bytes
View saved version →
---
name: video-dub
description: "A video re-voiced in another language: the speech translated slot by slot (or the user's own script), voices directed to match the original delivery, on-camera lines lip-synced, and the picture and music kept as they are. Use when the user wants an existing video's speech dubbed into another language with the picture kept as it is, from catalog voices or their own voice clips; not for subtitles, a new presenter video or remaking the video."
---
# Dub a video
A video re-voiced in another language: the speech translated slot by slot (or the user's own script), voices directed to match the original delivery, on-camera lines lip-synced, and the picture and music kept as they are.
Use this skill via the get_skill(skill_id="video-dub") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="video-dub", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
video-podcast888 Bytes
View saved version →
---
name: video-podcast
description: "A video podcast episode on a set: two hosts in a conversation, interview or debate (or one host solo), each with a fixed face and voice, lipsynced line by line, cut from singles and a two-shot. Use for a video podcast, talk show, interview or debate episode with on-camera hosts; for one person talking to camera without a podcast set, use talking-presenter, and for characters acting out a story, script-to-video."
---
# Video podcast
A video podcast episode on a set: two hosts in a conversation, interview or debate (or one host solo), each with a fixed face and voice, lipsynced line by line, cut from singles and a two-shot.
Use this skill via the get_skill(skill_id="video-podcast") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="video-podcast", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
video-prompting702 Bytes
View saved version →
---
name: video-prompting
description: "The limits and prompt format of each video model: which inputs it takes, how to cite references, time shots, write dialogue and phrase edits. Use before writing a prompt for generate_videos, and again when a video model refuses a request or a take comes back wrong."
---
# Video prompting guides
The limits and prompt format of each video model: which inputs it takes, how to cite references, time shots, write dialogue and phrase edits.
Use this skill via the get_skill(skill_id="video-prompting") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="video-prompting", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
video-recast1.02 KB
View saved version →
---
name: video-recast
description: "An existing video transformed through video-to-video: a person, outfit, prop, product or location swapped (Object Swap), or the whole cast and world rebuilt on the source's motion and camera (Motion Transfer), checked shot by shot. Use when a video the user already has should keep its exact motion, camera and timing while mapped people or things change: swap the presenter or the product, change an outfit, prop or location, or rebuild the whole look on the same moves. For a new cast speaking the lines shot by shot, use cast-swap."
---
# Recast a video
An existing video transformed through video-to-video: a person, outfit, prop, product or location swapped (Object Swap), or the whole cast and world rebuilt on the source's motion and camera (Motion Transfer), checked shot by shot.
Use this skill via the get_skill(skill_id="video-recast") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="video-recast", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
video-script1.25 KB
View saved version →
---
name: video-script
description: "The spoken lines and on-screen text for a short presenter or narrated video, written from a brief and laid out by video type (product demo, explainer, tutorial, sales pitch, announcement), with the spoken lines and the on-screen text block locked and a word count sized to the length. Use when the user wants the words of a video written, or a brief turned into them, before the video is made; it writes text only, and the talking-presenter, presenter-overlay, faceless-story-video or script-to-video skill then makes the video. Not for UGC ads (ugc-video writes its own) or podcast dialogue (video-podcast)."
---
# Video script
The spoken lines and on-screen text for a short presenter or narrated video, written from a brief and laid out by video type (product demo, explainer, tutorial, sales pitch, announcement), with the spoken lines and the on-screen text block locked and a word count sized to the length.
Use this skill via the get_skill(skill_id="video-script") tool call of this app: call it first; it returns the written steps for the video's words, as plain text. Its reference pages, also plain text, come from get_skill(skill_id="video-script", path=...) when the steps name them. Nothing in this skill is code to install or run.
Referenced files: 2 — file references are metadata, not package contents.
viral-effect-video1 KB
View saved version →
---
name: viral-effect-video
description: "A short effect video of a person or pet from one photo: a start frame built from the photo that keeps the face, then animated into a trend effect, a scene, a stunt, a transformation or a Halloween look. Use when the user wants their photo, or a frame of a short clip, turned into a viral effect, a 'me in a scene' or sports hype moment, a transformation, a seasonal look or a photo brought to life; to change their own footage while keeping its motion use video-recast, and for a person speaking a script talking-presenter."
---
# Viral effect video
A short effect video of a person or pet from one photo: a start frame built from the photo that keeps the face, then animated into a trend effect, a scene, a stunt, a transformation or a Halloween look.
Use this skill via the get_skill(skill_id="viral-effect-video") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="viral-effect-video", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
voiceover-timing976 Bytes
View saved version →
---
name: voiceover-timing
description: "Voiceover takes that fit fixed time windows, or one continuous narration that runs a set length: lines written to length, one locked voice, every take measured for length, pauses and pace, and only the misses rewritten. Use when narration has to fill set blocks or hit a total length, alone or from a narrated-video skill's voice step; for choosing a voice and delivery tags in general use audio-prompting, and for someone speaking on screen use talking-presenter."
---
# Voiceover timing
Voiceover takes that fit fixed time windows, or one continuous narration that runs a set length: lines written to length, one locked voice, every take measured for length, pauses and pace, and only the misses rewritten.
Use this skill via the get_skill(skill_id="voiceover-timing") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="voiceover-timing", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
voiceover-to-video1.07 KB
View saved version →
---
name: voiceover-to-video
description: "A storytelling video built from one narrator's voiceover: a planned story world, character sheets, location plates and scene anchors, then one clip per story beat, timed to the narration and joined under it. Use when the user has a voiceover, narration text or story script to be told by one off-screen narrator: a life story, a moral or folk tale, a Hindi or Hinglish kahani, a true-crime or history retelling. When only a topic is given and the narration still has to be written, use faceless-story-video. Not for characters speaking on screen (script-to-video), ads or a single clip."
---
# Voiceover to video
A storytelling video built from one narrator's voiceover: a planned story world, character sheets, location plates and scene anchors, then one clip per story beat, timed to the narration and joined under it.
Use this skill via the get_skill(skill_id="voiceover-to-video") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="voiceover-to-video", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
website-tour-ad930 Bytes
View saved version →
---
name: website-tour-ad
description: "A 9:16 creator-style ad for a website, web app or online store: a creator talks to camera the whole time while the user's own screenshots of the site pop up over them as cards, timed to the lines about them. Use when the site or app itself should be seen in a UGC-style video and the user can send screenshots of it; for a physical product shown in hand use ugc-video, and for a presenter with no site on screen use talking-presenter."
---
# Website tour ad
A 9:16 creator-style ad for a website, web app or online store: a creator talks to camera the whole time while the user's own screenshots of the site pop up over them as cards, timed to the lines about them.
Use this skill via the get_skill(skill_id="website-tour-ad") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="website-tour-ad", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.
what-if-shorts1.01 KB
View saved version →
---
name: what-if-shorts
description: "A 9:16 \"what if…\" short of about 60 s with one directed voiceover: a chosen look from 54 styles or the user's reference video, stills set in motion, a few generated clips, on-screen timers and counters, captions and an optional music bed; also as a series of parts ending on cliffhangers. Use for a scenario short that plays out a 'what would happen if…' question; for a narrated explainer, history or kids' story told over scenes use `faceless-story-video`."
---
# What-if shorts
A 9:16 "what if…" short of about 60 s with one directed voiceover: a chosen look from 54 styles or the user's reference video, stills set in motion, a few generated clips, on-screen timers and counters, captions and an optional music bed; also as a series of parts ending on cliffhangers.
Use this skill via the get_skill(skill_id="what-if-shorts") tool call: make it first and follow what it returns from the start. Its files are read with get_skill(skill_id="what-if-shorts", path=...) when it names them.
Referenced files: 2 — file references are metadata, not package contents.