Fuzzframe
OnlineNewNarrated videos made in Claude: motion graphics or illustrated stories, for YouTube, TikTok, Reels.
What it can do: 48 tools
guideread-onlyCall this FIRST, once per conversation. Returns whether the person's account can pay for the paid steps, the steps to guide them from an idea to a finished video, and the contract, kit and rules the film's code must follfuzzframe_actionRuns any of Fuzzframe's actions by name, including ones newer than the tools your chat knows (a chat keeps its own copy of a connector's tools for a while). action "list": every action now, each with what it does and itsplanread-onlyWhether the person's account has credits for the paid steps (recording the voice, starting a story, music, voice spaces, exports). Shows them a card with their video's look and title on top, then the plans, each openinglooksread-onlyThe looks a video can be made in. Shows the person a picker with a moving preview of each look (where their chat shows cards), and returns each look's palette, type, layout, motion and pacing, a preview frame for you, anlook_stillsread-onlyShows the person three looks with the video's own cover headline on each, in one card, with colours to try under the one they choose, "From a reference" and "See all". Nothing is rendered and nothing costs credits. Givechoose_lookSettles the look and colours of the video once they've chosen (or once you've chosen, in make it for me), remembers them for their next videos, and returns the look written out in those colours, with a colour of their ownotes_heardUsed by the review card: counts the notes tapped on a video that weren't saved as taste, and says when one has come back on another video in the same look, so it can be offered as taste.save_lookSaves a motion-graphics look of the person's own, made from a reference they shared (screenshots or frames of a video they like, a poster, a website) or described in their words, so it is offered first in the looks pickevoicesread-onlyThe person's saved voices, and the voice library searched by how the voice should sound (for example "calm even narrator", "warm british woman"). Shows the person a picker where they play each voice, pick one of their sasave_voiceSaves a voice to the person's voices (or removes it with saved false), so it shows first in the voice picker in every later chat and in Fuzzframe's own voice list. The picker's star does this by itself; call it when theybriefread-onlyIn make it for me (the guide's default), once the look is settled and you've written the script and picked the rest yourself: shows the person the whole brief on one card, in tabs (where their chat shows cards): for a mostart_storyFor an ILLUSTRATED STORY (Fuzzframe paints one picture per scene in one art style and a narrator tells it; the other kind, a motion film, you write as code): creates an illustrated animated short in the person's Fuzzframstory_looksread-onlyThe looks an illustrated story can be drawn in. Shows the person a picker with a sample picture of each look (where their chat shows cards). Free. Also list them yourself (name and one line each), for chats that don't shstoryWhere an illustrated story (from start_story) stands, and with go: true its next step: when the storyboard is planned it draws the pictures, records the voice and makes the film; when the film is made it exports the MP4.edit_storyChanges an illustrated story that's already made: the words of one scene, or what its picture shows. With no scene it lists the scenes, each with its number, words and what its picture shows. Nothing is drawn or recordedpost_textSaves the text to paste under an illustrated story when they post it, which shows in the studio's Post text tab (Fuzzframe writes none for a story: you do). Write it from the finished story: a line that adds to it, neverdescribe_lookPuts the art style of a picture the person uploaded (one of their own looks, from story_looks) into words, so every picture of a story can be drawn in it. Without description it shows you the picture; look at it and callplan_seriesSaves a series the person approved: its name, what it's about and for whom, what every episode is made with (kind, look, voice, length, tone, music, language) and the episodes in order (title, first line, angle, points).add_to_seriesPuts a video they already have, of either kind, into one of their series as its next episode, or moves it there from another series. The video itself doesn't change. Free. The video's id is the last part of its Fuzzframenext_episoderead-onlyA series' setup (kind, look, voice, length, tone, music), its taste, every episode with which are made, and the episode to make now (the next one not made, or the number asked). Call it when they ask for the next (or a gcontact_supportSends a message to the people who run Fuzzframe; one of them reads it on their phone and answers in the person's chat on fuzzframe.com (the chat button in the bottom corner of the app). Use it when the person asks for atasteread-onlyWhat this person likes in a video, learned from their own choices and reviews: their account's taste, and a series' taste when the video belongs to one (the series wins where both say something). Read it before planningremember_tasteSaves something about this person's taste, only when THEY ask for it to be remembered for a series or for all their videos ("always", "for this series", "from now on"); a note on one video is fixed in that video and notstart_videoCreates the video in the person's Fuzzframe account and records the voiceover once (this uses their credits: the brief's Make it is the OK in make it for me; in choose each step, picking the voice is the OK to record; avideoread-onlyWhere a video stands: the word timings, the film's files and whether they compile, the studio link to watch it, and the export's progress or MP4 link.save_storyboardShows the person the storyboard as a card to approve (a timeline on the recorded take, each beat with its time, words, on-screen plan and a sketch in the film's colours; once a plan is saved with save_animatic, a real rostageread-onlyThe instructions for one stage of making a film, with the rules that apply to it. Call it when the guide says, and follow what it returns for that stage.save_animaticSaves the film's storyboard as data (the animatic: beats on the recorded words, grid anchors, camera, transitions, events) with the look's tokens, and checks it. Every [fail] must be fixed and the plan saved again beforeqa_filmread-onlyRenders the film's keyframes from the saved plan and measures them: safe zones, overlaps, clipping, type sizes, contrast, the cover, and the code's timing, colours and easing. Each [fail] is objective: fix them all withsave_critiqueSaves the one technical look at the film's contact sheets (glitches, collisions code can't measure, the picture not matching the voice, a beat that doesn't do the motion its plan invented: kind flat). Once per film.write_fileWrites one file of the film (film/Film.tsx is the entry) and compiles the whole film. Returns what is wrong, or that it plays. Use edit_file for small changes afterwards.edit_fileReplaces one exact piece of text in a film file (it must occur exactly once) and compiles the film.read_fileread-onlyA film file as it is saved now.delete_fileRemoves a film file that is no longer used, and compiles the film.typecheckread-onlyChecks the film's TypeScript against the kit (wrong props, missing names, wrong types): the mistakes compiling alone doesn't catch. Run it before the first stills and after big changes.stillsread-onlyRenders frames of the film at the given seconds and returns them as one labelled sheet for you to LOOK at. Small (up to 12 frames, for the whole film) or large (up to 6, to read text and judge detail).picturesDraws pictures for a motion film on fal and links them to the video, so the film shows the thing itself (a person, a building, a product, an animal, a scene) and never a placeholder. kind "cutout": one subject on a transfal_modelsread-onlyFuzzframe runs its drawing, cut-outs, voices and music on fal.ai, and fal's whole catalogue is open to you: search it by words and/or a category (text-to-image, image-to-image, text-to-video, image-to-video, video-to-vidfal_runRuns any picture model from fal's catalogue (fal_models) for this video and keeps what it makes in the person's storage, on the film under a name: a drawing in another model, a photo restyled, an upscale, a background refind_picturesread-onlySearches Wikimedia Commons and Openverse (Flickr, museums, archives) for photos and artwork of a real thing the film names: a person, a place, a building, a product, an event, an animal. Only pictures a sold video may us
…and 8 more.
Connect it
claude mcp add --transport http fuzzframe https://fuzzframe.com/api/mcpSettings → Connectors → Add custom connector
URL: https://fuzzframe.com/api/mcp{
"mcpServers": {
"fuzzframe": {
"url": "https://fuzzframe.com/api/mcp"
}
}
}{
"servers": {
"fuzzframe": {
"type": "http",
"url": "https://fuzzframe.com/api/mcp"
}
}
}Questions
What does Fuzzframe do?
Narrated videos made in Claude: motion graphics or illustrated stories, for YouTube, TikTok, Reels. It exposes 48 tools to AI assistants.
What tools does Fuzzframe provide?
48 tools, including guide, fuzzframe_action, plan, looks, look_stills, choose_look, notes_heard, save_look, voices, save_voice, brief, start_story….
How do I connect Fuzzframe to Claude, ChatGPT or Cursor?
Add the remote MCP endpoint https://fuzzframe.com/api/mcp as a custom connector or MCP server. Copy-paste setup is on this page.
Does Fuzzframe require an account or API key?
No sign-in was needed to list its tools. Individual tools may still need an account or API key.
Is Fuzzframe online?
As of 2026-11-19 our probe found it online. It answered 100% of our last 3 checks.
Who publishes Fuzzframe?
fuzzframe.com. The official MCP Registry verified that the publisher controls this domain before accepting the listing.
Similar in Image & Video
-
Wixel by WixOnline · sign-in requiredOfficialPopularNew
Create images, logos, slide decks, videos and documents with Wixel, the AI design studio by Wix
-
Cloudinary Asset ManagementOnline · sign-in requiredOfficialPopularNew
Upload, generate, organize, search, and transform images, videos, and files with AI-powered tools.
-
ImageKit DevToolsOnlineOfficial
Official ImageKit developer server: search the docs and build transformation URLs. No sign-in.
-
ImageKit AdminOnline · sign-in requiredOfficial
Official ImageKit admin server: manage origins, URL endpoints, account usage and usage analytics.
-
ImageKit DAMOnline · sign-in requiredOfficial
Official ImageKit media library server: search, upload, organize, tag, share and manage assets.
-
JamOnline · sign-in requiredOfficial
Jam recordings (video, console, network) as context and proof of work for your coding agent.