Connect AI agents to ProductNow's context engine to search, create, review, and act.
ProductNow MCP Server (ai.productnow/productnow)
The ai.productnow/productnow Model Context Protocol (MCP) server connects AI agents to ProductNow’s context engine. It enables MCP clients to search, create, review, and act using a shared “living knowledge layer” provided by ProductNow’s cross-vendor storage layer. Clients authenticate with OAuth and operate as the signed-in ProductNow user.
🛠️ Key Features
Connects AI agents to ProductNow’s context engine
Supports MCP operations: search, create, review, and act
Exposes ProductNow’s knowledge layer to MCP clients
Uses OAuth authentication for the signed-in ProductNow user
🚀 Use Cases
Let AI agents query and work with ProductNow-stored, cross-vendor knowledge
Perform create/review/act workflows against the ProductNow context engine
⚡ Developer Benefits
Centralizes access to ProductNow’s knowledge layer via MCP
Consistent identity via OAuth as the signed-in ProductNow user
⚠️ Limitations
Provided as a public metadata and documentation package for marketplace submissions (documentation-only context)
Declared in the repository manifest (tools.json). Not verified against a live endpoint.
create_document
Create a new ProductNow document and immediately begin AI generation. Use `prompt` to specify the full document outline, including numbered sections, their titles, and their content. Use `context` for any additional supporting background — e.g., code snippets, meeting notes, etc. If the user references a file by name (e.g., 'the plan in swirling-sleeping-eclipse.md'), read that file first and pass its contents as `context`. Pass `templateVersionId` when the caller already has a template version to use.
Parameters
No parameters.
create_document_comment
Leave a comment on a document that is in review status. The comment is posted as the user's personal agent. Only works on review-status document versions. Use list_document_versions to find the correct documentVersionId first.
Parameters
No parameters.
create_folder
Create a new folder in ProductNow. Optionally nest it inside a parent folder. Returns the new folder's ID and name.
Parameters
No parameters.
create_help_document
Create a new blank ProductNow help document. Only available to ProductNow users. Returns the document ID, help page publication ID, and a URL to the document editor where content can be added.
Parameters
No parameters.
curate_knowledge_pack
Curate a shareable Knowledge pack page from a title, optional context and overview, ordered document items (one entry per document, with optional why-here notes and multiple highlight quotes each with a why-chosen reason), and optional further-reading documents. Filters out documents the caller cannot view, then returns a Knowledge pack URL. Write a short context blurb. Write summary as 3–5 short newline-separated bullet lines (prefix each with '- '), not a long paragraph. For each item note, write 2–4 sentences explaining what the reader will learn and when this doc matters — not a short caption. Prefer exact quotes as highlight text. For each highlight, include a one-sentence reason explaining why that quote was chosen. List each document once with a highlights array — never repeat the same documentId as separate items. Related must be distinct documents not already in items. Include 2–5 related docs outside the core pack when useful. Requires the knowledge_pack_application feature flag.
Parameters
No parameters.
get_document
Retrieve the full content of a ProductNow document by its ID, optionally targeting a specific version via `documentVersionId`. Returns the document name, version ID and status, folder location, and all section content. Use `search_knowledge_warehouse` first to find the document ID.
Parameters
No parameters.
get_document_chat
Fetch the current chat messages and edit mode for a document's AI draft chat. Returns the conversation history and whether the agent is in 'ask_before_edit' or 'edit_automatically' mode.
Parameters
No parameters.
get_document_feedback
Retrieve all feedback submitted on the latest review version of a document, returned as CSV. Columns: ID, Title, Description, Feedback, Feedback Type, Created On. Use search_knowledge_warehouse to find the document ID.
Parameters
No parameters.
get_document_threads
Fetch all comment threads on a specific document version. Returns thread IDs, statuses, quoted text, and creation timestamps. Use get_thread_messages to read the messages within a specific thread.
Parameters
No parameters.
get_thread_messages
Fetch all messages in a specific comment thread. Returns message IDs, content, author names, timestamps, and whether each message was posted by an AI agent. Use get_document_threads to find thread IDs first.
Parameters
No parameters.
import_prototype
Import a prototype from a published URL or from inline HTML/JSX/TSX file content. When sourceType is "url", provide a published prototype URL from an allowlisted platform (.lovable.app, .replit.app, .repl.co, .figma.site, .vusercontent.net, .vercel.app, .bolt.host, .created.app, .canva.site, .hostingersite.com). When sourceType is "html", "jsx", or "tsx", provide the raw file content as the source. Returns the hosted S3 URL and an iframe HTML snippet. Use this to convert a prototype into a resolved HTML file — the returned URL can be embedded in a prototype element.
Parameters
No parameters.
list_document_versions
List all versions of a document, optionally filtered by status ('draft', 'review', or 'published'). Returns version IDs, version numbers, statuses, and creation timestamps. Useful for locating a specific review or published version before fetching threads or commenting.
Parameters
No parameters.
list_folders
List folders and documents accessible to the current user, optionally filtered to a specific parent folder. Omit parentFolderId to list root-level contents. Use this to navigate the folder hierarchy and enumerate documents before creating documents or folders.
Parameters
No parameters.
list_help_documents
List ProductNow help documents accessible to the calling user. Optionally filter by category. Use get_document to read the full content of a specific help document.
Parameters
No parameters.
move_document_to_review
Move a ProductNow document's current version into REVIEW status. A fresh DRAFT version is created automatically to receive subsequent edits. Only DRAFT versions are moved; any other status (REVIEW, PUBLISHED, SNAPSHOT, ...) is a no-op — the response's `moved` flag and `previousStatus` report what happened. Targets the document's current version unless `documentVersionId` is provided.
Parameters
No parameters.
post_document_chat_message
Post a message to a document's AI draft chat to request edits or ask questions about the document. The message is sent to the draft agent which can modify the document's content, and works best when you target a specific document section to update. Use this to request changes to a document via natural language. Use the 'switch_document_chat_edit_mode' tool to set the edit mode of the chat to 'edit_automatically' to avoid requiring approval for your edits in the ProductNow UI.
Parameters
No parameters.
reply_to_thread
Reply to an existing comment thread on a document as the user's personal agent. Requires COMMENT permission on the document. Use get_document_threads to find thread IDs.
Parameters
No parameters.
search_folders
Find folders by name using fuzzy matching, ordered by relevance. Use this to resolve a folder name (e.g. "Salesforce") to its folderId before calling search_knowledge_warehouse with `destinationFolderId` to restrict a search to that folder and its subfolders.
Parameters
No parameters.
search_knowledge_warehouse
Search ProductNow documents and return answer-ready excerpts for factual questions. Pass query whenever the user names a topic, including when filters are also present. If the user asks only which documents match creator/date filters and names no topic, omit query to list documents by latest content edit; listing sources contain metadata but no excerpts. creatorNames matches document creators, not editors or contributors; its entries combine with OR, while creator and date filters combine with AND. Search workspace documents by default; use product_help only for questions about how ProductNow works. For a folder-scoped workspace search, resolve its id with search_folders and pass destinationFolderId; it is ignored for product_help. Topic-search results are relevance-filtered; all result sets are capped. Answer from the excerpts when possible, cite the documents you use, read a source document with `get_document` only when its excerpts are genuinely insufficient, and state uncertainty when evidence is weak or conflicting.
Parameters
No parameters.
switch_document_chat_edit_mode
Switch the edit mode of a document's AI draft chat. 'ask_before_edit' means the agent will ask for confirmation before making changes. 'edit_automatically' means the agent will apply edits directly without asking. Set to 'edit_automatically' to avoid requiring approval for your edits in the ProductNow UI.
Parameters
No parameters.
upload_media
Upload a media file (image or video) to ProductNow storage. Returns a signed backend `uploadUrl` valid for 5 minutes — upload the file via: `curl -X POST -F 'file=@<localPath>;type=<mimeType>' '<uploadUrl>'`. Also returns a `url` and `htmlText` you can embed in documents immediately.
ProductNow is the cross-vendor storage layer that joins your tools into one
shared context layer that AI can actually use. This hosted Model Context
Protocol (MCP) server makes that context layer available wherever you and your
team already work — through MCP clients that authenticate with OAuth and operate
as the signed-in ProductNow user.
This repository is a public metadata and documentation package for marketplace
submissions. It does not contain the ProductNow application source code or the
MCP server implementation.
ProductNow connects the tools where context already lives and turns that
signal into structured, queryable context — the living picture of your
organization. Through MCP, people and AI agents can search that company
brain, curate shareable knowledge packs, and collaborate on documents from
the clients they already use.
The MCP server exposes ProductNow workspace operations to compatible AI clients:
search the company brain for grounded, citeable evidence excerpts —
narrowed by folder, document creator, or last-edited date range, scoped to
ProductNow product help, or switched to folder-name search.
fetch a document (full text, sections, and comment threads) and
fetch_folder to browse the warehouse root, a folder, or a folder tree.
remember a fact, decision, or note and let ProductNow file it in the right
document and folder in the background.
Create documents with AI generation, edit_document through the document's
editing agent, poll get_status, and update_document_status to snapshot,
review, or publish.
Organize with create_folder, move, rename, and archive.
Collaborate with comment_on_document and reply_to_thread.
Allocate media upload slots with upload_media.
Curate shareable knowledge packs when the feature is enabled for the org.
Create ProductNow help documents when the caller is a ProductNow org member.
Note: The tool catalog in this repository was refreshed from the backend
MCP source on 2026-09-28. The live server's tools/list response remains the
runtime source of truth for connected clients. Some tools are gated by
organization membership or feature flags and may not appear for every user.
Quick Start
Use any MCP client that supports remote Streamable HTTP servers and OAuth
discovery.
When the client first connects, ProductNow returns an OAuth challenge that points
the client to public discovery metadata. The client should complete the browser
login flow and then send the issued bearer token to the MCP endpoint.
All tool calls run as the authenticated ProductNow user. ProductNow applies its
normal organization, workspace, document, and comment permissions to MCP
requests. Clients should treat write tools as consequential actions and ask for
user confirmation when appropriate.
