Skip to main content
Glama

Get a demo

get_demo
Read-onlyIdempotent

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, and fields must then list exactly the standard keys in form. 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 (see checklist.warnings). Other fields present in the document (newer features) are kept as they are; the server validates the whole document.

Requires: demos:read.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "demoId": "<demoId>"
    },
    "name": "get_demo"
  }
}

Input Schema

NameRequiredDescriptionDefault
demoIdYesDemo id (from list_demos), e.g. "d_Ab12Cd34Ef56Gh78".

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A3.5/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare read-only, idempotent, closed-world; the description adds what the annotations cannot — the 'Requires: demos:read' scope and a detailed account of what the response contains (screens, share settings, published state, hosted URL, embed code, revision, checklist). The strange clause about unknown newer fields being 'kept as they are' and server validation is muddled for a read tool, but the return-side disclosure is strong.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness2/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Roughly a thousand words that read as a full content-document reference manual (every tip/step field, narration clip internals, personalization tokens, checklist semantics) rather than a tool description. It is barely structured, front-loaded only by a single long paragraph, and most sentences concern the data model rather than invoking this tool.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema, documenting return structure has real value, and the description covers the document shape, auth scope, and follow-up tools comprehensively. It errs on the side of excess rather than omission, which is the safer failure mode here.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Single parameter with 100% schema description coverage, so baseline is 3. The description never mentions demoId, its format, or where to obtain it — that is left entirely to the schema, which does it adequately.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The opening clause — 'One demo in full: its content document ..., its screens ..., share settings, published state, hosted URL and embed code, and revision' — states a specific verb and resource with scope, and it distinguishes itself from get_screen ('Look at a screen with get_screen'). The purpose is unmistakable, but it is buried inside a single sprawling sentence rather than stated cleanly up front.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

There is genuine routing: 'Look at a screen with get_screen' and 'pass it [revision] to update_content'. But there is no explicit when-to-use / when-not statement, no contrast with siblings like get_guide or get_capture, and the routing hints appear as asides inside field documentation rather than as guidance.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.