MCP App Store
by KeenEthics
InterMIND icon

InterMIND MCP Server for Claude and ChatGPT

By MindMeeting OÜ

Overview

InterMIND is a space for everyday communication in the language you think in — meetings, chat and recaps, confidential and free for participants. Connect it once and Claude works with your InterMIND account from the conversation where it already works with your CRM, help desk or calendar. Meetings: create a meeting link you own (drop it into a calendar invite, a deal or a ticket; anyone who opens it knocks and you admit them) or schedule a meeting at a time, written to your connected Google or Microsoft calendar with the invitees. Reading: your recent and upcoming meetings, the newest recaps, a search over your recaps for what was decided, the transcript of a meeting by speaker, your channels and their messages, and the documents shared in your meetings. Acting: post a message in a channel as you — every member reads it in their own language. Ten tools: create_meeting_link, schedule_meeting and post_channel_message write; the other seven are read-only. OAuth 2.1 sign-in with your InterMIND account; the connection appears in Settings and can be revoked at any time. Claude only sees meetings you host or attend.

How to connect

  1. Claude: Open the InterMIND listing in the Claude connectors directory and choose Connect. Claude listing InterMIND exposes 2 tools here.
  2. ChatGPT: Open the InterMIND listing in the ChatGPT app directory and choose Connect. ChatGPT listing InterMIND exposes 23 tools here.

Tools

add_dlp_rules

ChatGPT
Add rules on what a chat message may not carry out of the organization — a pasted list, a table, a spreadsheet or a policy file the person attached or shared (a shared file is read with read_document) — and set the organization's say on the built-in rules for cloud and API keys, private keys, card numbers and IBANs (another action, or off). Use it when the person states what must not be sent in chat («contract numbers never leave», «IBANs only with a reason», «treat card numbers as audit only», «turn the AWS key rule off»). Each rule is a regular expression or a list of exact words with its own action: block, warn (send anyway with a justification, audited) or audit. The rules are run by a deterministic matcher on every message of the organization's meetings and channels, on the Business plan and above; no model decides at send time. Side effects: a rule with the same label replaces the earlier one, and every member's messages are checked from the next message on. Owner or admin of the organization only.

add_glossary_term

ChatGPT
Add one rule to the organization's translation glossary: how a word or phrase must be translated. «Translate X into French as Y» is term X, target_lang fr, rendering Y; «never translate X» or «X stays as it is» is term X with no rendering; «the recap rendered X as Y, it should be Z» is term X with rendering Z in the language of that recap. Use it when the person states how a term must or must not be translated. The translator applies the glossary to recaps and their transcripts, shared documents, chat and notes of accounts with AI features on, not to the live captions of a call. Side effects: the rule is shared by the whole organization, and an earlier rule for the same term and language is replaced. A chat or note message is recognised by the exact forms of the term (inflections are worked out automatically); «the translator missed X written as Y» is the same term again with Y in spellings. A list, a table or a file of terms is add_glossary_terms.

add_glossary_terms

ChatGPT
Add many rules to the organization's translation glossary in one call: a pasted list or table, a spreadsheet or CSV the person attached or shared (a shared file is read with read_document). Use it when the person hands over several terms at once; a single term is add_glossary_term. One rule per term and language, up to 200 per call; a longer list takes several calls. A sheet with the source term in one column and a language per further column is one rule per filled cell: the term, that column's language, the cell as the rendering (a cell equal to the source term means the term reads the same in that language). Each rule follows add_glossary_term: term, target_lang and rendering, or the term alone to keep it as written in every language. Side effects: the rules are shared by the whole organization, an earlier rule for the same term and language is replaced, a duplicate in the list keeps the last one; a batch that would exceed the organization's cap is refused whole and nothing is written.

get_channel_messages

ChatGPT
Read one page of a channel's chat, oldest first: sender, time, the text as written and, when the room already translated it, the text in the connected account's language. Use it when the person wants to see what was written in a channel; older pages are read with before.

get_transcript

ChatGPT
Read the latest finalized transcript of a meeting the connected account hosts or attends, as Markdown grouped by speaker, in the language it was spoken (or the reader's cached translation). Use it when the person wants the actual words of a meeting; a summary is list_recaps. Long transcripts are cut at max_chars with truncated: true; the transcript_url opens the whole text.

list_channels

ChatGPT
List the team chats (channels) and personal inboxes the connected account belongs to, by last activity. Use it to find a channel's id for get_channel_messages, post_channel_message or read_team_page.

list_dlp_rules

ChatGPT
List the organization's data protection rules as the matcher runs them on chat messages: the built-in rules (keys, private keys, card numbers, IBANs) with the organization's action or off mark, then the organization's own rules with their id, kind, action and source — a file given to the assistant or the owner's word on a message in the chat. enabled says whether the plan runs them (Business and above). Use it when the person asks what is protected, why a message was refused, or needs a rule's id to remove it.

list_documents

ChatGPT
List the files shared in a meeting or channel the connected account belongs to — uploads, linked Google Drive or OneDrive files and files sent in the chat (a spreadsheet, a PDF, a Word file) — newest first, each with the id read_document takes. Use it when the person asks which files were shared; without meeting_id it covers the account's recent meetings.

list_glossary_terms

ChatGPT
List the organization's translation glossary as the translator reads it: every term with its rendering per language or the «keep as written» mark, 40 terms a page. Use it when the person asks what terms the glossary holds, to check a rule before changing it, or with query to find the id of one rule for remove_glossary_term (a narrowed list carries the ids).

list_meetings

ChatGPT
List meetings the connected account hosts or attends: recent by last activity (default) or upcoming (scheduled in the future, soonest first), each with its link and its recap page. Use it when the person asks what meetings they have, or to find a meeting's id for get_transcript, list_documents or search_knowledge.

list_recaps

ChatGPT
List the newest recaps of meetings the connected account hosts or attends, newest first: title, when the meeting happened, links to the recap and transcript pages, and the summary in Markdown. Use it when the person asks what their recent meetings were about; a specific topic is search_knowledge or search_recaps.

list_team_pages

ChatGPT
List the organization's team status pages, computed from its meeting recaps — today one page: the open tasks and promises by assignee, with the due date when one was said. Use it before read_team_page to learn which pages exist, whether pages are on for the account and which windows (7, 28 or 90 days) are available.

post_channel_message

ChatGPT
Post a chat message into a channel as the connected account. Use it when the person asks to write, send or post something in a channel. Side effects: every member sees the message immediately, translated into their own language, and it becomes part of the channel's recap; this connection cannot edit or delete it afterwards. Returns the message id and the channel link.

read_document

ChatGPT
Read the full extracted text of a file shared in a meeting or channel the connected account belongs to, by the document id from list_documents or from a document passage of search_knowledge (a file name is not an id). Use it when the person needs the content of a shared file. Long texts are cut at max_chars with truncated: true.

read_team_page

ChatGPT
Read one team status page of the organization: rows are facts extracted from meeting recaps, each with the verbatim quote it rests on, the link to the recap it comes from and its effective verdict (open, done, not done); an open task stays until judged and is marked stalled after 28 days without movement. Use it when the person asks what is open, who owes what or what the team promised. Slices: the whole organization (default), one channel (channel_id), or the channels sharing a tag (tags, set by the host in the channel's settings). The reader sees only what their membership allows: an admin the whole organization, a member the sessions they hosted or attended plus rows assigned to them. Pages come from list_team_pages; a row is judged with set_team_task_verdict.

remove_dlp_rule

ChatGPT
Remove one data protection rule of the organization by its id or its label, or turn a built-in rule off by its id («seed:iban»). Use it when the person asks to drop a rule («forget the rule about contract numbers», «IBANs may be sent again»). Side effects: from the next message on, nothing is matched by that rule for any member; a removed rule stays in the audit trail. Owner or admin of the organization only.

remove_glossary_term

ChatGPT
Remove rules from the organization's translation glossary. Use it when the person asks to drop a term («forget the rule about X», «X may be translated again»). The term as written removes every rule of that term in every language; a rule id from list_glossary_terms removes that one rule. Side effects: the removal is immediate and permanent for the whole organization, and the translator stops applying the rule; a term that is not in the glossary is an error.

schedule_meeting

ChatGPT
Create an InterMIND meeting at a given local time and, when the connected account has a Google or Microsoft calendar connected, write it into that calendar with the invitees, who receive the calendar invitation. Use it when the person names a clock time (tomorrow at 10:00); a link without a time is create_meeting_link. Returns the link, the UTC window and whether the calendar event was created; without a connected calendar the link still works but carries no time.

search_knowledge

ChatGPT
Search everything the connected account can see in InterMIND — recaps, transcripts, chats and the text of shared files — by meaning and by words at once. Returns the best-matching passages, each with the meeting, the date and the file it comes from. Use it for questions about what was said, decided, sent or written in the account's meetings and chats; narrow with meeting_id, a date range or the sources when the question names them. Questions about the InterMIND product itself are search_product_docs.

search_product_docs

ChatGPT
Search InterMIND's own product documentation: features, plans, settings, integrations and how things work. Use it for questions about the InterMIND product itself; the account's own meetings, chats and files are searched with search_knowledge.

search_recaps

ChatGPT
Find recaps whose title or summary mentions given words (word forms match: pricing finds price), over meetings the connected account hosts or attends, newest first. Each hit is a snippet with the matches in bold plus links to the full recap and transcript. Use it with one or two distinctive words (a name, a topic) when the person wants the recaps about something; a question in natural language is search_knowledge.

set_team_task_verdict

ChatGPT
Record the connected member's verdict on a task or promise of a team status page: done closes the row, not_done keeps it open with a short reason. Use it when the person says a task or promise is done or not done. Side effects: the member's verdict replaces the verdict the model gave, a done row leaves the next recap's carried list, and every verdict is written to the audit trail; a later call records a new verdict over it. Only rows the member may read, by the id from read_team_page.

list_recaps

Claude

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/intermind"><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/intermind)