Connect ChatGPT, Gemini & CLI Agents to Your Data Rooms
Plug Web AI chats or terminal agents into ConeShare over remote MCP. Group multiple AI conversations into separate Data Rooms to manage and share independently, mint watermarked links for existing files, and keep everything on your own infrastructure.
Add https://app.coneshare.com/mcp/sse as a custom MCP server in ChatGPT, Gemini, or Claude.ai.
Walkthrough
See ChatGPT + ConeShare MCP in action
Sign in to ConeShare from ChatGPT to create protected links for your files, or group multiple AI conversations into separate Data Rooms to manage and share independently.
1Sign in to authorize ChatGPT
Paste the ConeShare MCP URL and approve read-only or read-write access directly in your browser.
2Create protected document links
Find existing files from chat and mint links with email gates, passwords, expiry, or PDF watermarks.
3Organize & share chats by Data Room
Save multiple AI conversations into separate Data Rooms, manage permissions per project, and share each room independently.
Setup
Connect in under a minute
No local server or daemon required. Sign in to authorize browser-based AI apps, or paste a scoped API key into CLI and desktop clients.
Remote MCP Server URL
https://app.coneshare.com/mcp/sse1Paste the remote MCP URL
In ChatGPT, Google Gemini, or Claude.ai, add a custom MCP connector pointing to the URL above (or your self-hosted domain).
2Sign in & authorize AI access
Sign in to ConeShare in the browser popup and choose read-only or read-write access.
3Start prompting in chat
Ask your AI to create a share link for an existing file or save your conversation as a trackable ConeShare document.
If you self-host, replace app.coneshare.com with your own domain.
- Approve what your AI can access
- When connecting a Web AI client, you sign in to ConeShare and choose read-only or read-write access. Destructive deletes, admin endpoints, and viewer PII stay blocked by default.
- Server-enforced link controls
- Email verification, passwords, NDA gates, expiration dates, and dynamic watermarks (PDF, Office, and images) are enforced server-side regardless of the prompt.
- Self-hosted & storage-agnostic
- Keep your files in Nextcloud, Google Drive, Dropbox, or your own infrastructure. Saved AI memos and deal documents never leave your control.
MCP Tools
29 tools across documents, Data Rooms, and links
Everything your Web AI or CLI agent needs to save AI outputs, search Data Rooms, issue protected links, and inspect engagement.
Documents & Folders
12 toolscreate_text_document()Save AI-generated Markdown or plain text (chat summaries, research memos) as a document
list_documents()Paginated list of workspace documents with folder filtering
get_document()Retrieve document metadata, versions, and active links
search_documents()Search documents by title or description
update_document()Rename or update metadata of an existing document
delete_document()API key onlySoft-delete a document (moves to Trash)
create_folder()Create a folder in the workspace hierarchy
update_folder()Rename an existing workspace folder
delete_folder()API key onlySoft-delete a workspace folder
move_items()Move documents or subfolders into destination folders
request_document_upload()API key onlyRequest a pre-signed URL for direct storage uploads
finalize_document_upload()API key onlyCommit uploaded document metadata and start preview rendering
Web AI clients only receive the read-only or read-write access (read / write) you approve when signing in to ConeShare. Destructive deletes, file uploads, viewer analytics, and admin tools require a scoped API key.
FAQ
Frequently Asked Questions
Add https://app.coneshare.com/mcp/sse (or your self-hosted domain) as a custom MCP server in your AI client, then sign in to ConeShare and approve read-only or read-write access in your browser.
Ready to connect your AI to ConeShare?
Start on ConeShare Cloud or self-host on your own server, and connect your AI clients in seconds.