MCP App Store
by KeenEthics
StrataSpace icon

StrataSpace MCP Server for ChatGPT

By StrataSpace, Inc.

Overview

Work with your Strata documents in ChatGPT. Find documents across your accessible workspace, read and summarize their contents, and make targeted edits or create new documents. Open the embedded Strata editor to work with documents and interactive HTML previews. Export presentations as audience-ready HTML or PDF. Sign in to Strata to access your documents; actions use your authorized permissions. PDF presentation exports consume the document owner’s capture allowance.

How to connect

  1. ChatGPT: Open the StrataSpace listing in the ChatGPT app directory and choose Connect. ChatGPT listing StrataSpace exposes 34 tools here.

Tools

accept_suggestion

ChatGPT
Accept a pending suggestion, applying its edit to the document.

app_document_history

ChatGPT
Editor app only. Lists a document's edit history, or the block-level changes in one history entry, for the embedded editor's history panel. Read-only.

app_get_section_content

ChatGPT
Get the full content of a document section by document ID and section ID. Returns the section text and heading context; records a document-view audit event.

app_get_ws_token

ChatGPT
Editor app only. Issues a short-lived WebSocket token so the embedded document editor can join live collaboration on a document. Records a document-view audit event.

app_list_comments

ChatGPT
Editor app only. Lists a document's comment threads for the embedded editor's comment panel. Read-only.

app_manage_comment

ChatGPT
Editor app only. Creates, replies to, edits, deletes, resolves, reopens, or reacts to comments on a document from the embedded editor's comment panel. Side effect: a created comment or reply that @mentions an agent starts a task for that agent on the document.

app_presentation_validation_allowance

ChatGPT
Read the presentation owner's capture balance before validation. Uncached whole-deck counts consume an inspection rate-limit token and store a cached slide count; capture allowance is unchanged.

app_resolve_image_urls

ChatGPT
Editor app only. Returns short-lived download URLs for the images embedded in a document so the embedded editor can display them. Read-only.

app_restore_document_version

ChatGPT
Editor app only. Restores a document to an earlier version from its edit history, replacing the current content for every collaborator. The restore is recorded as a new history entry, so it can itself be undone while that history is retained.

create_comment

ChatGPT
Comment on a document: start a thread, optionally anchored to quoted text, or reply to one with parentCommentId. Side effect: a comment or reply that @mentions an agent starts a task for that agent on the document. Mention an agent only when the user asked for that agent to act.

create_document

ChatGPT
Create a document from inline Markdown. folderId sets the audience: a tier folder is visible to its members; omit it for the owner-only personal root. For content that already exists as a file, use presign_document_upload instead. HTML BLOCKS Code blocks with language html render as live sandboxed previews: no cookies, storage, or network. The only loadable URLs are the vetted libraries /sandbox/libs/mermaid.min.js (then mermaid.initialize({startOnLoad:false}); mermaid.run()), /sandbox/libs/d3.min.js (global d3), /sandbox/libs/three.module.min.js (ES module), and /sandbox/libs/doom.js. The app's CSS variables follow light and dark mode and are a safe default: --color-foreground, --color-muted-foreground, --color-border and --color-card for text and surfaces, --color-primary (brand cyan) and --color-secondary-foreground (brand purple) for emphasis, --color-success, --color-warning and --color-destructive for status, and --color-chart-1 to --color-chart-5 for data series. You are free to use your own palette, gradients or imagery instead, especially when the request asks for something creative. Either way, use color to encode meaning rather than one accent on grey, say what each color means, and if you set your own colors, set background and text together so both modes stay readable. Height starts near 360px and follows the content. To chart data from the same document, name a json/csv/tsv/yaml fence (``json name=sales), a table or image with <!-- strata:name=sales --> on the line above it, or use the reserved name frontmatter (scalar fields). The html fence declares up to 8 names (``html uses="sales,team"). Inside the preview strata.data.get(name) returns {kind:'text', format, text} (parse it yourself; d3.csvParse, not split(',')), {kind:'table', columns, rows} (cells are strings), {kind:'frontmatter', values}, {kind:'image', url, width, height, alt}, or {kind:'unavailable', reason}. Re-render on the window strata:data event. BLOG FRONTMATTER A blog: mapping in the first YAML frontmatter block designs the published post. typography: modern | editorial | expressive | technical. palette: strata | paper. blockControls, tableOfContents, coverInBody: visible | hidden. readingSurface: opaque | translucent (translucent only over a color or gradient wallpaper). accent, background: {light: "#RRGGBB", dark: "#RRGGBB"}, quoted; body text and accent need 4.5:1 contrast in both schemes. fonts.heading, fonts.body: Manrope | Syne | Fira Code | Georgia | Times New Roman | Arial | Helvetica | Verdana | Trebuchet MS | Palatino | Courier New | IBM Plex Mono. summary: up to 160 chars. tags: up to 10, each up to 32 chars. noindex: true | false. series: {name (up to 80 chars), order (integer from 1)}. cover, hero: the name of an image in this document, set with <!-- strata:name=cover --> on the line above it. wallpaper: exactly one of color {light, dark}; gradient {type: linear | radial, stops: 2 to 5 {light, dark}}; pattern {name: dots | grid | lines | paper | topographic | stars, tint: {light, dark}}; image (image name); html (name of an html block, plus fallback: {color: {light, dark}} or {image: name}). Unknown keys, malformed YAML, or invalid values block publishing; edits report blogValidation findings. A presentation: map whose settings do not parse turns presentation mode off; edits, and read_document by a caller who can edit, then report presentationSettings with the parser message.

create_document_from_template

ChatGPT
Create a document from a template, substituting its declared variables. Templates live in templates folders and declare name, description, and variables in frontmatter.

export_presentation

ChatGPT
Export a presentation as standalone HTML or printed PDF. Returns a download URL valid for fifteen minutes. Presenter notes are removed from slide decks; opaque whole-deck HTML is rejected because note removal cannot be proven. A draft whose presentation: settings do not parse is refused: a caller who can edit the document gets the parser message, any other reader a refusal without it. A published edition always exports, with default settings in place of any that do not parse. PDF consumes the document owner's capture allowance; HTML is unmetered. PDF prints every slide in one headless page under a single sixty-second render deadline, with WebGL software-rendered and animation frozen to a still once a small render budget is spent, so a deck of many heavy scenes can time out where the HTML export succeeds.

finalize_document_upload

ChatGPT
Finish an upload after its PUT, creating the document or replacing an existing one's content. Idempotent. Returns { documentId, bytes, created }.

find

ChatGPT
Search, list, or discover documents, spaces, and folders across the workspace. intent selects the kind of discovery (search, list, recent, related, info) and scope selects where to look. intent='info' with scope.type='document' returns a body-free summary — documentVersion, open comment and pending suggestion counts, a link, and presence when the collaboration group is enabled — the cheap way to poll whether a document moved since you read it. List folders and their scopes (a document's audience comes from its folder) with intent='list', scope.type='folders'. scope is a JSON object with a type field, not a string. Examples: {"scope":{"type":"all"}}, {"scope":{"type":"space","id":"space_..."}}, {"scope":{"type":"document","id":"doc_..."}}. Search uses hybrid keyword + semantic retrieval fused via Reciprocal Rank Fusion, so both short keywords and natural sentences work as queries. Filters cover status, owner, date ranges, includeShared, direction, sinceDays (for intent='recent'), and tokenBudget. Returns a uniform shape with items that carry id, title, snippet, score, and source attribution. Narrow scopes can return zero results even when matching content exists elsewhere in the workspace.

get_company_theme

ChatGPT
Read company theme CSS and its verified source version for copying into a presentation.

get_document_graph

ChatGPT
Get the connection graph neighborhood around a document. Returns nodes (documents) and edges (connections) discovered by BFS traversal. Optional parameters: - depth: BFS depth (default 2, max 5) - maxNodes: Maximum nodes to return (default 50, max 200)

get_image

ChatGPT
Retrieve an image from a document as base64-encoded data. The imageId is found in document content as [Image: alt text (imageId: <id>)] markers or ![alt](strata://image/<id>) in markdown. Returns the image as inline base64 PNG data that can be viewed directly.

insert_document_section

ChatGPT
Insert Markdown blocks at a heading boundary without touching existing content: position is {kind: beforeSection | afterSection, sectionPath} or {kind: endOfDocument}. sectionPath is a heading path such as "Architecture > Database". Markdown follows the HTML BLOCKS and BLOG FRONTMATTER rules in create_document.

list_comments

ChatGPT
List comment threads on a document, optionally only active or resolved ones. truncated is true when more threads exist than were returned.

list_company_themes

ChatGPT
List company presentation themes available to copy.

list_suggestions

ChatGPT
List pending suggestions on a document.

move_document

ChatGPT
Move a document into a folder, which changes who can see it. A personal folder is owner-only, a tier folder is visible to its members; to make a document private, move it to the user's personal folder.

presign_document_upload

ChatGPT
Start uploading a Markdown file into a new or existing document. Avoids re-emitting bytes you already have. Returns { uploadId, uploadUrl, requiredHeaders, expiresAt }. Upload sequence: 1. presign_document_upload with contentLength (the file's byte count) and documentId to replace, or title [+folderId] to create 2. PUT the file bytes to uploadUrl with every requiredHeaders entry 3. finalize_document_upload with uploadId Markdown follows the HTML BLOCKS and BLOG FRONTMATTER rules in create_document.

propose_suggestion

ChatGPT
Propose a tracked edit that replaces exact text in a document, for a person to accept or reject. Refresh proposals carry baseVersion, sourceRef {documentId, sourceVersion}, and anchor {sectionPath, blockOrdinal, offset}.

read_document

ChatGPT
Read a document's content by ID. Returns rendered Markdown. Records a document-view audit event. Large documents automatically return a structural outline instead of full content (default budget: 10000 tokens). The sections parameter accepts heading paths from that outline and returns only those sections.

reject_suggestion

ChatGPT
Reject a pending suggestion without applying it.

replace_document_text

ChatGPT
Replace exact text in a document with new Markdown; set replaceAll to change every match. Prefer this targeted edit over rewrite_document. Markdown follows the HTML BLOCKS and BLOG FRONTMATTER rules in create_document.

rewrite_document

ChatGPT
Replace a document's whole body with inline Markdown. For content that already exists as a file, use presign_document_upload instead. Markdown follows the HTML BLOCKS and BLOG FRONTMATTER rules in create_document.

set_comment_resolved

ChatGPT
Resolve a comment thread, or reopen a resolved one with resolved false.

share_document

ChatGPT
Grant a person access to a document by email, emailing an invitation to an external recipient when needed.

unshare_document

ChatGPT
Revoke a user's access to a document.

validate_presentation

ChatGPT
Render presentation slides and return layout, contrast, binding, structure, and theme findings. Always executes checks and stores captures. Costs one capture per slide per viewport, with no repeat charge for the same render key this cycle. Checks at most two slides at one viewport per call. While status is more, the result carries a next continuation that covers the remaining slides and viewports. When status is failed, reason says why: renderTimedOut (a slide did not finish in time; check one slide per call or simplify its heaviest drawing), renderFailed, rendererUnavailable, conflict (safe to run again), limitReached, invalidRequest or internal. Omitted selection covers all slides at both viewports.

withdraw_suggestion

ChatGPT
Withdraw your own pending suggestion.

Capabilities

WritesInteractive

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/strataspace"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/strataspace)