AtriumKit
Server Details
Edit, translate, narrate, publish and analyze AtriumKit interactive demos from any MCP client.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- OAuth
- Works in Glama
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- Repository
- atriumkit/atriumkit-mcp
- GitHub Stars
- 0
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 50 tools
The set has many distinct resources, but several tools overlap: add_scene vs write_scene both add scenes, update_scene vs write_scene both modify them, and update_content vs update_step/add_step/delete_step/move_step cover the same step data at different granularities. Descriptions clarify most cases, but an agent can still misselect among the scene and step editing tools.
All names are snake_case and mostly follow verb_noun (create_demo, get_video, update_step). A few use prepositional forms (add_to_collection, replace_in_copy) and the demo/video voiceover pair is similar, but the convention is predictable overall.
50 tools is heavy for an agent-facing surface, with many granular operations split across demos, steps, videos, scenes, brand kits, captures, collections, and analytics. While the platform is broad, the count exceeds a practical well-scoped set and encourages fragmentation.
Core demo and video workflows are well covered (create/list/get/update/publish, steps, scenes, voiceovers, analytics), but major lifecycle gaps remain: no delete_demo or delete_video, no create/update/remove for collections beyond add_to_collection, and no create/delete for brand kits or captures. These gaps will force agents to work around missing operations.
Publisher details
- Operator
- Simplisik LLC, the company that makes AtriumKit · Publisher source
- Operator website
- https://www.atriumkit.com
- Vendor relationship
- First-party
- Privacy policy
- Not provided
- Trust center
- Not available
- Restrictions
- Requires an AtriumKit account (the Free plan works, no admin approval). Translations, voiceovers, personalized links and knowledge base search need the Pro or Team plan, and AI rewrites and translations use the workspace's AI credits. · Publisher source
Available Tools
50 toolsadd_sceneAdd a sceneAInspect
Adds a scene from a motion-kit block (list_video_blocks) with values for its variables, at a position (default: the end). Images are the video's assets: values like "asset:". Up to 40 scenes.
Requires: demos:read, demos:write.
Inputs: block (required), videoId (required), label, intent, +4 more.
add_stepAdd a stepAInspect
Adds a step to a guided demo: a tooltip ("tip", pointing at spot on a screen), a chapter card, a call to action ("cta", a button link), a lead form ("form"), an intro or an outro. Placed after after (a step id; "start" for first), else at the end (before a final outro). Unset fields get the editor’s defaults; tooltips use the screen of the step they follow unless you pass screenId. fields sets anything else on the step.
Requires: demos:read, demos:write.
Inputs: kind (required), demoId (required), cta, body, +7 more.
add_to_collectionAdd a demo to a collectionAIdempotentInspect
Adds a demo to a collection’s list (at the end, or at 0-based position), optionally with a title and description shown there instead of the demo’s own. Only live demos appear to viewers. Adding a demo that is already in it changes nothing.
Requires: demos:read, demos:write.
Inputs: demoId (required), collectionId (required), title, position, +1 more.
create_demoCreate a demo from web pages or screenshotsAInspect
Makes a new draft demo without the Chrome extension, from either:
urls: 1–20 public web pages (http or https). AtriumKit opens each in its own headless browser (desktop 1440×900 or a 390 px phone), closes an obvious cookie banner, blurs emails and phone numbers, and captures a screenshot plus the page’s HTML (editable like a recording). Each page becomes a screen, in order. Pages behind a sign-in are refused ("needs a sign-in"): record those with the extension. Capturing takes about 5–10 seconds a page; this tool waits up towaitseconds, then answers with the progress.images: 1–30 screenshots, each an httpsurlof a PNG/JPEG/WebP or base64data(the whole request must stay under 2 MB, so prefer URLs for anything but small images). Guided demos get an intro, one tooltip per screen and an outro; clickable demos link the pages through their own navigation. Then, unlessai.copyis false, AI writes the copy (and, withai.voice, narration and voice) in the given tone and language towardgoal(AI credits; voice is Pro and Team). Answers with demoId, editUrl, the capture and AI progress, and what to do next (get_demo_progress while it runs). The demo is a draft: publish_demo puts it live.
Requires: demos:read, demos:write.
Inputs: ai, type, urls, wait, +4 more.
create_personalized_linksCreate personalized linksAInspect
Makes personalized links for a demo, one per prospect: each link's values fill the demo's variables for whoever opens it ({{company}}, {{first_name}}… in step copy; built-ins first_name, last_name, full_name, company, title, email, domain, plus the custom ones in content.variables). Each link can name its rep ({{rep.first_name}}, {{rep.calendar}}…) and an intro video. Up to 500 at once, e.g. from a CRM list. The URL carries only an opaque code (?p=…); the values are stored with the link as workspace data, like leads. Values for variables the demo doesn't have are dropped. Pro and Team plans. Returns each link's hosted URL and embed URL.
Requires: demos:read, demos:write.
Inputs: links (required), demoId (required).
create_videoCreate a blank videoAInspect
A blank video (one title scene) to build by hand with add_scene. To make a video from a demo or a brief, use generate_video.
Requires: demos:read, demos:write.
Inputs: title, aspect, themeId.
delete_sceneDelete a sceneADestructiveIdempotentInspect
Removes a scene from the video (undo: the studio’s undo, or the chat’s Undo).
Requires: demos:read, demos:write.
Inputs: sceneId (required), videoId (required).
delete_stepDelete a stepADestructiveIdempotentInspect
Removes one step from a guided demo’s draft. Choices and "next" links that pointed at it go where it led (its own "next": a step or the end), like in the editor; when it had no "next" they are cleared (a choice becomes unlinked, a "next" goes back to the following step). Revision history keeps the previous version (restore it in the dashboard’s History).
Requires: demos:read, demos:write.
Inputs: demoId (required), stepId (required).
duplicate_demoDuplicate a demoAInspect
Copies a demo (screens, content, settings, translations) into a new draft named " (copy)". Useful for a variant per audience.
Requires: demos:read, demos:write.
Inputs: demoId (required).
generate_videoGenerate a video (beta)AInspect
Beta. The AI director makes a motion video from a brief and/or up to 3 demos (their screens, copy and narration): it picks a look, writes a storyboard of 10–15 beats (hook, product moments, CTA), gathers screens, b-roll, music and sound effects, designs every scene and records the narration. It runs as a job (usually 1–3 minutes): this waits up to 20 seconds, then answers with the finished video or its progress — if it is still "generating", call get_video later instead of generating again. Uses AI credits (5 up to a minute, 10 longer; refunded if it fails) and AI voice characters for narration. A walkthrough keeps a demo step's recorded narration when its script stays word for word; other formats write their own script. On the Free plan the user's email must be verified first, and the brand kit styles one video: the first made from the workspace's first demo (list_brand_kits brandVideo); others get the built-in look.
Requires: demos:read, demos:write.
Inputs: look, media, music, proof, +7 more.
generate_video_voiceoversGenerate narrationAIdempotentInspect
Records AI narration for scenes whose script has no current audio (default: all of them). Uses AI voice characters; scene durations should fit the narration (get_video shows them).
Requires: demos:read, demos:write.
Inputs: videoId (required), sceneIds.
generate_voiceoversGenerate narration (voiceovers)ADestructiveInspect
Gives a demo spoken narration: for each step (guided) or screen (clickable) a short script read by an AI voice when it appears, with captions. Pass scripts to set them (saved to the draft first), writeMissing to draft scripts with AI where there are none (AI credits), and optionally voice (list_voices). Then every target whose clip is missing or out of date gets a new clip; unchanged ones are skipped and identical inputs reuse earlier clips for free. Uses the workspace’s monthly AI narration characters (Pro and Team, active subscriptions): it stops when they run out and reports what was done. The clips go live with publish_demo. Translations’ narration is generated in the editor.
Requires: demos:read, demos:write.
Inputs: demoId (required), voice, scripts, targets, +2 more.
get_analyticsGet analyticsARead-onlyIdempotentInspect
Views, unique visitors, completion rate, average time and leads for the last days (default 30, today included, UTC), with daily counts and the change against the previous period. For one demo (demoId): the step funnel (sessions that reached each published step) with where viewers drop off and the biggest drop, screen visits and hotspot clicks (clickable), recent sessions and traffic sources. Linear demos: the drop is from the previous step. Branched demos (branched: true, choices or step jumps): steps on different paths can’t be compared in list order, so the drop is left, the sessions that stopped at that step without finishing, with each step’s path, plus paths and a breakdown per choice. Without demoId: the workspace totals and top demos. Free plans get view counts only (detailed: false).
Requires: demos:read, analytics:read.
Inputs: days, demoId.
get_brand_kitGet a brand kitARead-onlyIdempotentInspect
One brand kit in full: its settings (accent, tooltip surface, font, corner radius, brand name, logo, header button), the kit document (kit: colors, fonts, buttons, forms, CTAs, hosted page; null values are derived), the colors viewers actually get (resolved, with contrast warnings) and its knowledge kit (product, audience, voice, glossary terms) that AI copy uses.
Requires: demos:read.
Inputs: themeId (required).
get_captureGet a library captureARead-onlyIdempotentInspect
One capture of the library and where it is used: every demo (id, title, guided or clickable, draft or live, whether it has unpublished changes) and each of its screens that links to the capture (id, name, the capture version it shows, whether it is still updating, how many steps or edits a replacement asked to look at, how many edits of its own the demo made on it). Use get_demo and get_screen for the screens themselves.
Requires: demos:read.
Inputs: captureId (required).
get_demoGet a demoARead-onlyIdempotentInspect
One demo in full: its content document (guided: steps; clickable: hotspots and startScreenId; plus description), its screens (ids, names, sizes, captured page title and address, whether it has an HTML snapshot, and livePageChecked on live pages viewers can scroll past the screenshot: whether someone has scrolled through it in the editor), share settings, published state, hosted URL and embed code, and revision (pass it to update_content). When the demo has flows, checklist lists them with their sizes and what to fix. Look at a screen with get_screen.
Steps (guided demos), in order:
tip: { id, kind:"tip", screenId, title ≤200, body ≤2000, spot:{x,y,w,h} (percent of the screenshot), place:"auto"|"top"|"right"|"bottom"|"left", advance:"click"|"next"|"timer"|"narration" ("narration": continues when the step's narration ends), timerMs 1000–60000, hotspot:"pulse"|"dot"|"outline"|"hidden" ("pulse": the spot with a pulsing ball on its top-right corner; "dot": only the pulsing ball, in the spot's middle, the spot still the click target), zoom, dim, camera?, cursor?, spotlight? } Tooltip camera and motion (optional; left out = the demo default in settings.motion): "camera": { mode?: "auto" (zoom to the spot) | "frame" (show camera.frame) | "off", frame?: {x,y,w,h} percent of the screenshot (fitted to the stage, 1–4× zoom), transition?: "glide"|"cut", speed?: "slow"|"normal"|"fast" }, "cursor": true|false (animated cursor that moves to the spot and clicks), "spotlight": "rounded"|"circle" (the cutout shape with dim). Without camera.mode, zoom: true means "auto".
intro / chapter / cta / outro cards: { id, kind, screenId (or null), title, body, cta ≤60 (button label), ctaUrl (https://, mailto: or tel:, or null), theme:"dark"|"light" }
form (lead form): { id, kind:"form", screenId, title, body, cta, fields: ["email", "name"?, "company"?, "phone"?] (email required), form?, theme }.
form(optional) lists every field in order, custom questions included: { key, type: text|textarea|email|phone|number|url|select|radio|checkboxes|consent|hidden, label, placeholder, required, options? (choices), param? (hidden: the URL parameter) }; the standard ones use keys email/name/company/phone with label null for the default, andfieldsmust then list exactly the standard keys inform. Branching (any guided step but a form): "choices": 2–4 of { id, label ≤60, target: {type:"step",stepId} | {type:"end"} | {type:"url",url} | null } (a card shows them instead of its button; a tooltip shows them as buttons) — set_choices is the easy way. "next" on any step: {type:"step",stepId} or {type:"end"}: where the default path goes after it (absent: the following step); use it to skip ahead, rejoin or end a path. Personalization: titles, bodies, button labels, choice labels, hotspot labels and links can hold {{key}} or {{key|fallback}} (built-ins first_name, last_name, full_name, company, title, email; custom ones declared in content.variables as { key, label, default?, sample? }). Each viewer sees their own values (personalized links, URL parameters), everyone else the fallback or default. Keep the tokens when rewriting copy. Hotspots (clickable demos): { id, screenId, rect:{x,y,w,h}, label ≤120, hint ≤300, target: {type:"screen",screenId} | {type:"url",url} | null }, plus "auto": true on navigation links made by stitch_demo (shown quietly) and "link": a group id the same control shares on every screen. Plus startScreenId and description (≤300, shown under the title on the hosted page). Narration (voiceovers, Pro and Team): "voiceover" on any guided step: { script ≤1000 (what a voice says when the step appears: short, spoken, separate from the tooltip text), source:"ai"|"recorded"|"uploaded", clip: { key, durationMs, cues, hash, voiceId? } | null }. Clickable demos keep it per screen in "screenVoiceovers": { "": same }. "voice": { voiceId (list_voices), speed 0.7–1.2, stability 0–1 } is the demo's AI voice. Make clips with generate_voiceovers; never write a clip key yourself (it must be this demo's audio). Checklist (a demo with several flows, every plan): "flows": { checklist: true|false (viewers see it), title? ≤60 (else "Checklist"), flows: up to 12 of { id, title ≤60, steps: [step ids] (guided) or [screen ids] (clickable) } }. A step or screen is in one flow at most; a guided flow usually starts at a chapter card and runs to the next one, or is the path a start choice opens. Viewers see % complete, steps and time per flow and done marks, and can jump to a flow. Ids of deleted steps are left out (seechecklist.warnings). Other fields present in the document (newer features) are kept as they are; the server validates the whole document.
Requires: demos:read.
Inputs: demoId (required).
get_demo_progressCheck a new demo’s progressARead-onlyIdempotentInspect
How a demo made with create_demo is coming along: the pages being captured (each one’s status, title or why it failed), whether the demo exists yet (demoReady), and the AI pass writing its copy and voice (per part: copy, narration scripts, voice clips; notes say why a part was skipped, e.g. no AI credits left or voice not on this plan). Works for any demo’s latest AI pass too.
Requires: demos:read.
Inputs: demoId (required).
get_guideGet a guideARead-onlyIdempotentInspect
A published demo as a step-by-step guide, exactly as viewers get it (the published version: no draft changes, presenter notes or private sources): Markdown with a picture of each step (the screen with the clicked spot), plus an outline of sections and items (step ids, numbers, titles, picture URLs, branching choices). In any language the demo is published in (locale; default its source language). Every plan; the demo must be published (publish_demo).
Requires: demos:read.
Inputs: demoId (required), locale.
get_screenLook at a screenARead-onlyIdempotentInspect
A screen of a demo as viewers see it: the screenshot (blurs and text edits applied, scaled to at most 1568 px) plus the readable text on it with positions. HTML screens add an outline of their elements (ref, tag, text, rect). Rects are percent of the screenshot, the same units as a tooltip step’s spot and a hotspot’s rect, so you can point a step at an element. Hidden, blurred and personal data never appears.
Requires: demos:read.
Inputs: demoId (required), screenId (required).
get_videoGet a videoARead-onlyIdempotentInspect
A video: its look, voice, music, captions and scenes in order. Each scene has a label, what it is for (intent), the kit block it came from (null: custom), start and duration (ms), transition, narration with its audio status, and its editable variables with current values (change them with update_scene). While the AI director is still making it, generation says how far it got. With includeHtml, each scene’s HTML too (only needed to hand-edit a scene with update_scene html; to change how a scene looks or moves, use write_scene).
Requires: demos:read.
Inputs: videoId (required), includeHtml.
get_video_usageVideo export allowanceARead-onlyIdempotentInspect
Seconds of exported video left this month, and whether exports carry the "Made with AtriumKit" mark.
Requires: demos:read.
Inputs: no named parameters.
list_brand_kitsList brand kitsARead-onlyIdempotentInspect
The workspace’s brand kits (themes: how demos, cards, forms and hosted pages look): id, name, whether it is the default (new demos and videos use it), accent and primary colors, fonts, brand name, logo, header button, and whether it has a knowledge kit. editable says whether the plan lets you change them (Pro: one kit, Team: up to 20; Free can only rename, and viewers see its kit on the workspace’s first demo only, firstDemo, and on the first video made from it, brandVideo). A demo picks its kit with update_demo { themeId }.
Requires: demos:read.
Inputs: no named parameters.
list_capturesList library capturesARead-onlyIdempotentInspect
The workspace’s capture library: captured screens (a screenshot and, with HTML capture, the page) that screens in many demos link to, so one edit or replacement updates every demo. Each has an id, name, page title and address (without query or ids), size, whether it has an HTML page, its version and how many demos and screens use it. Most recently changed first. Linked screens show libraryCaptureId in get_demo.
Requires: demos:read.
Inputs: kind, query, usage.
list_collectionsList collectionsARead-onlyIdempotentInspect
The workspace’s collections (demo centers: one hosted page grouping several demos): id, title, link, access, how many demos are on it and live, and the demo ids in order.
Requires: demos:read.
Inputs: no named parameters.
list_demosList demosARead-onlyIdempotentInspect
Lists the workspace’s interactive demos, most recently updated first: id, title, type (guided = tooltip walkthrough, clickable = hotspot sandbox), status (draft or live), step/screen counts, views, hosted URL when live. Filter by title search, folder, status or type. Paginate with cursor.
Requires: demos:read.
Inputs: type, limit, query, cursor, +2 more.
list_leadsList leadsARead-onlyIdempotentInspect
People who left their details in a demo’s lead form or email gate (or a collection’s gate), newest first: email, name, company, phone, which demo, source, country, referrer, time, the answers to the form’s custom fields (answers: key, label, value), the campaign they came from (utm: source, medium, campaign, term, content, and ad click ids) and their intent score (0–100: how engaged their viewing session was; null when not computed). Search by query (email, name or company), filter by demo or by utmSource / utmCampaign (exact values), sort: "intent" for the most engaged first. Paginate with cursor. This is personal data: use it for the user’s request only.
Requires: demos:read, leads:read.
Inputs: sort, limit, query, cursor, +3 more.
list_personalized_linksList personalized linksARead-onlyIdempotentInspect
The personalized links of a demo, newest first: label, hosted URL and embed URL, views, first and last view. The stored values (prospect names, companies, emails) are personal data and come back only with the leads:read scope. Without it, labels made from the values show as "Link ", and query matches the labels shown on each page only (pagination ignores the query). Paginate with cursor.
Requires: demos:read.
Inputs: demoId (required), limit, query, cursor.
list_video_blocksList scene blocksARead-onlyIdempotentInspect
The motion kit's ready-made scenes (add_scene fills one with values; write_scene can start from one) and its looks (presets for update_video), plus the authoring rules hand-written scene HTML must follow. Looks: editorial (White canvas, big refined type, pastel marker blocks behind words, words blurring into focus. Trust, enterprise, AI products, thoughtful launches.); bold (Electric color field, heavy tight display type that pops in, yellow highlights, punchy cuts. Launches, teasers, social ads.); terminal (Near-black with a faint grid, monospace headings that type in, green tokens, glowing hairlines. Developer tools, APIs, CLIs, infrastructure.); cinematic (Warm paper with film grain and a drifting light leak, quiet light type, slow fades, footage in big rounded frames. Brand films, lifestyle, storytelling.); playful (A sunset sky with drifting clouds and floating shapes, chunky outlined cards, bouncy words, hand-drawn circles. Consumer, SMB, friendly products.); product (Crisp UI on a soft brand gradient, the camera and cursor doing the talking. Walkthroughs, feature tours, onboarding.).
Requires: demos:read.
Inputs: no named parameters.
list_videosList videosARead-onlyIdempotentInspect
The workspace’s videos (AI-directed motion videos made from demos or a brief), most recently changed first: id, title, status, aspect, length, scene count, whether the latest export matches.
Requires: demos:read.
Inputs: query.
list_voicesList narration voicesARead-onlyIdempotentInspect
The AI voices narration can use: ElevenLabs default voices (they speak every language) and a few from its voice library per language, with id, name, gender, accent, age, language and description. Also the workspace’s AI narration status: whether AI voices are available (Pro and Team: an active subscription, or a trial with a smaller allowance; recording and uploading happen in the editor), the voice model, and this month’s character allowance. Pass a voice id to generate_voiceovers.
Requires: demos:read.
Inputs: language.
list_workspacesList workspacesARead-onlyIdempotentInspect
The AtriumKit workspace this connection works in (one per connection), with your role, its plan, AI credits, folders and brand kits (themes). Use the folder and theme ids with list_demos and update_demo.
Requires: demos:read.
Inputs: no named parameters.
move_sceneMove a sceneAIdempotentInspect
Moves a scene to another position (0 = first).
Requires: demos:read, demos:write.
Inputs: sceneId (required), videoId (required), position (required).
move_stepMove a stepAIdempotentInspect
Moves a step of a guided demo: to a 0-based toIndex, or right after / before another step id.
Requires: demos:read, demos:write.
Inputs: demoId (required), stepId (required), after, before, +1 more.
publish_demoPublish a demoADestructiveInspect
Publishes the current draft: viewers of the hosted link, embeds and collections see it right away (a new version, vN, kept in revision history). Fails with the reason when something needs fixing first (no steps, a broken link) or the plan’s published-demo limit is reached. Returns the hosted URL and embed code, and uncheckedLivePages: live pages viewers can scroll further than their screenshot that nobody has scrolled through in the editor yet (tell the user; they open them from editUrl).
Requires: demos:read, demos:publish.
Inputs: demoId (required).
replace_in_copyFind and replace in the copyADestructiveInspect
Replaces a word or phrase everywhere in a demo’s copy, in one save: guided steps (titles, bodies, button labels, choice labels), clickable hotspots (labels and hover hints), the description, personalization defaults and guide overrides; narration scripts too with includeNarration (their clips then need generate_voiceovers again). {{variables}} are never touched. Case-insensitive unless matchCase; wholeWord skips matches inside longer words. dryRun only counts and lists the places. Not the demo title (update_demo) and not the text on the screens themselves (screenshots and HTML pages: the demo editor’s Find & replace). Saved to the draft; revision history keeps the old copy.
Requires: demos:read, demos:write.
Inputs: find (required), demoId (required), replace (required), dryRun, +3 more.
replace_video_textFind and replace in a videoADestructiveInspect
Replaces a word or phrase across a whole video in one save — e.g. renaming a company or product everywhere ("Globex" → "Acme"): every scene's text variables (headlines, callouts, buttons, list items, labels…, including copy still at its default), scene labels, the narration scripts with includeNarration (default true; their audio then needs generate_video_voiceovers), and the video title. Case-insensitive unless matchCase; wholeWord skips matches inside longer words; dryRun only counts and lists. Images, colors and asset keys are never touched. Words typed into a custom scene's markup instead of a variable can't be replaced here: the answer lists those scenes (rewrite them with write_scene).
Requires: demos:read, demos:write.
Inputs: find (required), replace (required), videoId (required), dryRun, +5 more.
rewrite_copyRewrite copy with AIADestructiveInspect
Rewrites a demo’s copy with AtriumKit’s AI (the same strings as “Write with AI” in the editor), following instruction, e.g. "for a CFO audience, shorter, mention month-end close". Covers every step (guided: titles, bodies, button labels, choice labels; a card with choices has no button) or hotspot (clickable: labels, hints), or just stepIds / fields; optionally the demo title and description. {{variables}} in the copy are kept. Navigation links made by stitching ("auto") are left alone unless you name them in stepIds; then only their label is rewritten, once per link group, so the same control reads the same on every screen. With apply (default) the result is saved to the draft (revision history keeps the old copy); otherwise you get suggestions to review. Uses the workspace’s AI credits (1 per 20 strings).
Requires: demos:read, demos:write.
Inputs: demoId (required), instruction (required), apply, fields, +3 more.
score_demoScore a demoARead-onlyIdempotentInspect
A 0–100 quality score for a demo with a short list of fixes, worst first: length and pacing, copy, a clear call to action, personalization, narration, branching, how it works on phones, lead capture, and broken or stale screens. Each fix says what to do and links to where in the editor. Free; with ai: true the AI also reads the copy and rates its clarity (1 AI credit, editors only).
Requires: demos:read.
Inputs: demoId (required), ai.
search_knowledgeSearch the knowledge baseARead-onlyIdempotentInspect
Searches the workspace’s internal knowledge base: published guides and their steps (every published language), knowledge kits (product summary, audience, notes, links, glossary terms) and passages of uploaded knowledge files (sales decks, price sheets, docs). Ranked results, each with its kind (guide, step, product, glossary, file), a snippet and where it is; open a guide or step with get_guide. Pro and Team plans.
Requires: demos:read.
Inputs: query (required), limit.
set_choicesSet a step’s choices (branching)ADestructiveInspect
Turns a tooltip or card of a guided demo into a fork: 2–4 buttons ("I'm in finance" → step X, "Book a call" → a link, "Skip" → end). A card shows them instead of its single button; a tooltip shows them as buttons (clicking its spot still follows the default path). Each choice goes to a step id, the end of the demo, or a URL; leave goTo out to link it later (viewers then continue on the default path). Pass an empty list to remove the choices. To make a path skip ahead or rejoin, set "next" on its last step with update_step, e.g. { "next": { "type": "step", "stepId": "st_…" } } or { "next": { "type": "end" } }. The result lists the paths and anything to fix (unreachable steps, paths that never end).
Requires: demos:read, demos:write.
Inputs: demoId (required), stepId (required), choices (required).
stitch_demoLink navigation across screens (clickable)AIdempotentInspect
Makes every sidebar item, tab and menu link of a clickable demo work from every screen it shows on, like the real app, not only where the recording clicked it: each control links to the recorded screen it leads to (by its link address, or because it is the same control as a recorded click). Run it with apply: false first to see what it adds ("links" across "screens", per control in "groups") and the weaker matches it only suggests ("suggestions", each with a key and why); then apply: true, optionally with accept (suggestion keys) to add some of those too. New recordings are linked this way automatically; use this for demos recorded before, or after adding screens. Links it adds are quiet in the player (no pulse) and grouped per control. It never changes or duplicates existing hotspots.
Requires: demos:read, demos:write.
Inputs: demoId (required), apply, accept.
translate_demoTranslate a demo with AIAInspect
Adds a language to a demo (if it isn’t there yet) and translates its copy with AI: the missing and outdated strings, or all of them. Viewers get the language picker and ?lang= after the next publish_demo. Pro and Team plans; uses AI credits (1 per 50 strings).
Requires: demos:read, demos:write.
Inputs: demoId (required), locale (required), retranslate.
unpublish_demoUnpublish a demoADestructiveIdempotentInspect
Takes a demo offline: its hosted link and embeds stop showing it. The draft and its history stay; publish_demo brings it back.
Requires: demos:read, demos:publish.
Inputs: demoId (required).
update_brand_kitEdit a brand kitADestructiveIdempotentInspect
Changes a brand kit: name, accent, kit colors (null derives one again from the accent and contrast), tooltip surface, heading and body fonts (built-in or any Google Fonts family; files are self-hosted on save), corner radius, brand name, the hosted page’s header button, or makes it the workspace default. Fields you leave out stay. Every demo that uses the kit renders with the change, published ones included. Pro and Team (Free can only rename).
Requires: demos:read, demos:write.
Inputs: themeId (required), name, accent, colors, +8 more.
update_contentEdit the content documentADestructiveInspect
Saves a demo’s content document (steps or hotspots, startScreenId, description, the checklist’s flows) to the draft: either the whole document (content) or JSON Patch operations (ops, RFC 6902, paths like "/steps/2/body" or "/hotspots/0/label"; "-" appends). Validated like the editor (ids unique, screens exist, safe links) and recorded in revision history. Pass the revision you read with get_demo to avoid overwriting someone else’s edit. For single steps, update_step / add_step / delete_step / move_step are simpler.
Steps (guided demos), in order:
tip: { id, kind:"tip", screenId, title ≤200, body ≤2000, spot:{x,y,w,h} (percent of the screenshot), place:"auto"|"top"|"right"|"bottom"|"left", advance:"click"|"next"|"timer"|"narration" ("narration": continues when the step's narration ends), timerMs 1000–60000, hotspot:"pulse"|"dot"|"outline"|"hidden" ("pulse": the spot with a pulsing ball on its top-right corner; "dot": only the pulsing ball, in the spot's middle, the spot still the click target), zoom, dim, camera?, cursor?, spotlight? } Tooltip camera and motion (optional; left out = the demo default in settings.motion): "camera": { mode?: "auto" (zoom to the spot) | "frame" (show camera.frame) | "off", frame?: {x,y,w,h} percent of the screenshot (fitted to the stage, 1–4× zoom), transition?: "glide"|"cut", speed?: "slow"|"normal"|"fast" }, "cursor": true|false (animated cursor that moves to the spot and clicks), "spotlight": "rounded"|"circle" (the cutout shape with dim). Without camera.mode, zoom: true means "auto".
intro / chapter / cta / outro cards: { id, kind, screenId (or null), title, body, cta ≤60 (button label), ctaUrl (https://, mailto: or tel:, or null), theme:"dark"|"light" }
form (lead form): { id, kind:"form", screenId, title, body, cta, fields: ["email", "name"?, "company"?, "phone"?] (email required), form?, theme }.
form(optional) lists every field in order, custom questions included: { key, type: text|textarea|email|phone|number|url|select|radio|checkboxes|consent|hidden, label, placeholder, required, options? (choices), param? (hidden: the URL parameter) }; the standard ones use keys email/name/company/phone with label null for the default, andfieldsmust then list exactly the standard keys inform. Branching (any guided step but a form): "choices": 2–4 of { id, label ≤60, target: {type:"step",stepId} | {type:"end"} | {type:"url",url} | null } (a card shows them instead of its button; a tooltip shows them as buttons) — set_choices is the easy way. "next" on any step: {type:"step",stepId} or {type:"end"}: where the default path goes after it (absent: the following step); use it to skip ahead, rejoin or end a path. Personalization: titles, bodies, button labels, choice labels, hotspot labels and links can hold {{key}} or {{key|fallback}} (built-ins first_name, last_name, full_name, company, title, email; custom ones declared in content.variables as { key, label, default?, sample? }). Each viewer sees their own values (personalized links, URL parameters), everyone else the fallback or default. Keep the tokens when rewriting copy. Hotspots (clickable demos): { id, screenId, rect:{x,y,w,h}, label ≤120, hint ≤300, target: {type:"screen",screenId} | {type:"url",url} | null }, plus "auto": true on navigation links made by stitch_demo (shown quietly) and "link": a group id the same control shares on every screen. Plus startScreenId and description (≤300, shown under the title on the hosted page). Narration (voiceovers, Pro and Team): "voiceover" on any guided step: { script ≤1000 (what a voice says when the step appears: short, spoken, separate from the tooltip text), source:"ai"|"recorded"|"uploaded", clip: { key, durationMs, cues, hash, voiceId? } | null }. Clickable demos keep it per screen in "screenVoiceovers": { "": same }. "voice": { voiceId (list_voices), speed 0.7–1.2, stability 0–1 } is the demo's AI voice. Make clips with generate_voiceovers; never write a clip key yourself (it must be this demo's audio). Checklist (a demo with several flows, every plan): "flows": { checklist: true|false (viewers see it), title? ≤60 (else "Checklist"), flows: up to 12 of { id, title ≤60, steps: [step ids] (guided) or [screen ids] (clickable) } }. A step or screen is in one flow at most; a guided flow usually starts at a chapter card and runs to the next one, or is the path a start choice opens. Viewers see % complete, steps and time per flow and done marks, and can jump to a flow. Ids of deleted steps are left out (seechecklist.warnings). Other fields present in the document (newer features) are kept as they are; the server validates the whole document.
Requires: demos:read, demos:write.
Inputs: demoId (required), ops, content, revision.
update_demoUpdate demo detailsADestructiveIdempotentInspect
Changes a demo’s title, description (shown under the title on the hosted page), share settings, brand kit (themeId from list_workspaces), folder, access or link (slug: /d/). Saved to the draft and to revision history; title, description and theme reach viewers on the next publish_demo, settings, access and the link apply at once (earlier links keep working). Changing access also needs demos:publish.
Requires: demos:read, demos:write.
Inputs: demoId (required), slug, title, access, +4 more.
update_sceneUpdate a sceneAIdempotentInspect
Changes one scene directly (free, instant): its variable values (merged; the ids and current values are in get_video — text, colors, numbers, images as "asset:"), label, intent, duration (600–30000 ms; keep it at least as long as its narration), transition into it, narration script (null removes it), or its whole HTML (≤ 60000 characters, following the authoring rules from list_video_blocks; its variables are re-read from it). Changing the script makes its narration stale: then call generate_video_voiceovers. To change how a scene looks or moves, ask write_scene instead of writing HTML.
Requires: demos:read, demos:write.
Inputs: sceneId (required), videoId (required), html, label, +5 more.
update_stepEdit a stepADestructiveIdempotentInspect
Changes fields of one step of a guided demo, e.g. { "title": "…", "body": "…" }, a tooltip’s "spot" or "place", a card’s "cta" and "ctaUrl", "next" (where the default path goes after it). Fields you leave out stay as they are; null removes an optional field: { "next": null } goes back to the following step, { "choices": null } removes the choices, and the same for "notes", "camera", "cursor" and "spotlight" ("ctaUrl": null just removes the link). id and kind can’t change. Text can use {{variables}} (see get_demo). Saved to the draft (revision history keeps the old version).
Requires: demos:read, demos:write.
Inputs: demoId (required), stepId (required), changes (required).
update_videoUpdate a videoAIdempotentInspect
The video’s title, aspect, look (preset, colors, fonts, motif, pace, radius), captions, voice and music level. Only the fields you send change. Changing the look restyles every scene (they use its CSS variables).
Requires: demos:read, demos:write.
Inputs: videoId (required), look, music, title, +3 more.
write_sceneWrite a scene with AIAInspect
The AI director writes one scene from an instruction, in the video's look. With sceneId it rewrites that scene — its layout, motion, style or content ("make it a before/after", "show the dashboard filling up", "punchier, with the numbers counting up") — keeping what you don't ask to change; without sceneId it adds a new scene at position (default: the end), optionally starting from a kit block (list_video_blocks). The scene is checked against the authoring rules and fixed automatically before it is saved. Uses 1 AI credit per call (refunded if it fails). For plain copy, timing or narration changes use update_scene instead (free and instant).
Requires: demos:read, demos:write.
Inputs: videoId (required), instruction (required), block, sceneId, +1 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Build, edit and update narrated, interactive training courses from your documents.
Write, rehearse, record and publish website demo videos, from the desktop app or the cloud.
Find, edit, publish and export the interactive product demos in your HowdyGo library.
Turn a URL and one line of intent into a narrated product demo video, from any MCP client.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceEnables agents to create, edit, preview, and publish presentation decks with themed slides, audience Q&A and voting, imports, and post-talk results through a hosted MCP endpoint.MIT
- AlicenseAqualityBmaintenanceEnables MCP clients to operate local Adobe Premiere Pro projects, covering media ingest, bins, sequence and timeline editing, transitions, effects, keyframes, captions, markers, metadata, proxies, multicam, color, audio, and exports through 283 discoverable tools, 13 context resources, and 10 guided prompts. A bundled CEP panel provides the supported production bridge, with read-only connection verification and higher-level workflows for assembling product spots and motion-graphics demos.5MIT

Cueflow MCP Serverofficial
AlicenseNot gradedqualityBmaintenanceEnables AI assistants to author, validate, and preview narrated, animated UI walkthroughs from static HTML scene kits, producing self-contained multilingual HTML guides without recording video.45 npm2MIT- FlicenseNot gradedqualityBmaintenanceEnables publishing and managing authenticated HTML artifacts and creating credential-free questionnaires with signed public answer links through a unified MCP service.-
Glama MCP Gateway
Add one secure layer between your agents and this server.