MCP App Store
by KeenEthics
Ideals icon

Ideals MCP Server for Claude and ChatGPT

By Ideals

Overview

Run your Ideals data room from Claude – build project structures, manage users, search and summarise documents, track project activity, and more – all through natural conversation, no compromises on security, and without switching tools

How to connect

  1. Claude: Open the Ideals listing in the Claude connectors directory and choose Connect. Claude listing Ideals exposes 50 tools here.
  2. ChatGPT: Open the Ideals listing in the ChatGPT app directory and choose Connect. ChatGPT listing Ideals exposes 53 tools here.

Tools

accept_terms_of_use

ChatGPT
Accept Terms of Use. Records the current user's acceptance of a project's Terms of Use, using the agreementId returned by get_terms_of_use. Only call after the user has explicitly agreed to the terms. After acceptance, retry the action that required it.

add_users_to_project

ChatGPT
Invite participants. Adds new users to a project and assigns them to a group; always include securitySettings.accessFrom with the current date. The new user inherits the group's document permissions. All security settings are ignored if project is archived.

answer_question

ChatGPT
Answer a question. Submit an answer to a Q&A question. The backend automatically determines the post type and target post based on the caller's Q&A role. Recommended workflow: 1. Call get_qna_configuration to determine the user's Q&A role. 2. Call get_qna_threads to find threads in an answerable status. 3. Call this endpoint with the thread ID and answer body. Role-based behavior (automatic — the caller only provides `body`): - Expert — creates an AnswerDraft. Thread must be in Assigned status, and the user must be the assigned expert. - Answer Coordinator — creates an Answer. Thread can be in Submitted, Assigned, Proposed, or Rejected status. If an Approver is configured, the answer is held for approval. - Answer Approver — creates an Answer (acts as approval). Thread must be in ToApprove status. - Question Submitter / Question Drafter — cannot use this endpoint.

ask_question

ChatGPT
Ask a question. Create a new Q&A question in the specified project. Recommended workflow: 1. Call get_qna_configuration first to fetch available question teams, categories, and participant roles. 2. Pick a teamId from the returned questionTeams[] list and, optionally, a categoryId from categories[]. 3. Call this endpoint with the chosen values. For a detailed, role-aware guided workflow, use the AskQuestion MCP prompt resource (when supported by your client).

assign_experts

ChatGPT
Assign experts. Assign experts to one or more Q&A threads (up to 100 per request). The expertIds list replaces the current experts of every listed thread: experts not in the list are unassigned. To add an expert while keeping the current ones, include the current experts as well. To unassign all experts from the threads, send an empty expertIds list (or omit the field). Recommended workflow: 1. Call get_qna_configuration to get the expert user IDs from experts. 2. Call search_qna_threads to find the threads and their current assignees. 3. Call this endpoint with the thread IDs and the full list of experts each thread should have. Who can assign: - Answer Coordinator — threads in Submitted, Assigned, Proposed or Rejected status. - Expert — only threads assigned to them in Assigned or Proposed status, and only when expertsCanReassignQuestions is true in get_qna_configuration. Experts cannot unassign all experts: an empty expertIds list returns 403 Forbidden. Each thread is processed independently. Threads that cannot be assigned are returned with status NotAssigned, an errorCode and an errorMessage; the other threads are still assigned. Duplicate thread IDs are processed once. The optional comment is posted to each thread whose experts change and sent to the newly assigned experts. Threads that already have exactly the requested experts are reported as Assigned but left unchanged: no comment is posted and no notification is sent, so repeating the request is safe. Returns 400 Bad Request when any expert ID does not belong to an active expert of the project; no thread is changed. Returns 403 Forbidden when the caller's Q&A role cannot assign experts (including an Expert sending an empty expertIds list) or the project is locked. Returns 404 Not Found with error code 3009 if the Q&A module is not enabled for this project.

cancel_closure_schedule

ChatGPT
Cancel scheduled closure. Use when asked to keep a project open, prevent it from closing, or undo a scheduled closure.

close_project

ChatGPT
Close project. Permanently closes a data room or schedules it to close at a specific date/time, terminating access for ALL participants including administrators. Works from any project status. Destructive and irreversible once executed – use only when a project is signed/ended or access must be cut. Before calling, ask a confirmation from a user to execute this.

copy_documents

ChatGPT
Copy documents. Copy one or more files and/or folders to a destination folder within the same project. The originals remain in place. > [!NOTE] > Items are processed sequentially. If an item fails (for example, due to insufficient permissions), processing stops and the response contains only the items that succeeded before the failure. Required scope (integrations): documents.write

create_folder

ChatGPT
Create new folder. Creates a single folder in a project. The building block for setting up or extending a data-room structure (folder index). For a full multi-folder index, call repeatedly.

create_folders

ChatGPT
Create folders. Creates one or more folders in a project in a single call. Input: - id (optional): GUID that uniquely identifies the folder within this batch. Only needed when another item in the same batch will reference it via parentId; if omitted, one is generated automatically. When you do supply one, generate a random UUID v4 (e.g. "550e8400-e29b-41d4-a716-446655440000") as if produced by Guid.NewGuid() — never reuse or hardcode values. - name: display name for the folder - parentId (optional): GUID of the parent folder; must reference either an existing folder in the project (the base location) or another id within the same batch. Omit to place the folder at the project root. Nesting: Forward references are supported: a folder can reference a parentId that appears later in the same batch, allowing an entire deep hierarchy to be submitted in one call without ordering constraints. Response: Returns a summary with the total count of successfully created and failed folders, plus per-item details indicating success or failure reason for each entry. Notes: - Up to 500 folders can be created in a single request; larger batches are rejected. - If the project or a referenced base folder does not exist or is inaccessible, or you lack folder-creation permission, the whole request is rejected and nothing is created. - Duplicate names under the same parent are created with an automatic suffix (e.g. "Reports (2)"). For a single-folder request the returned name reflects the final stored name; for a multi-folder batch, the response echoes the requested name even if it was suffixed. - A folder with an invalid name is reported as a failed item (its name is not corrected) and its in-batch child folders are skipped; other valid folders are still created. - If VDR reports a batch failure without indicating which item(s) failed, the affected items are reported with an error asking you to verify their outcome manually rather than assumed successful. - Prefer batching over repeated single-folder calls to minimize round-trips.

create_group

ChatGPT
Create new group. Create new group with the specific role, feature permissions, document permissions, and security settings. Required scope (integrations): users.write

create_project

ChatGPT
Create new project. Creates a new project (data room) under the corporate account. Follow up with folder, group, and invite tools to fully set up the room.

delete_documents

ChatGPT
Move documents to recycle bin. Move one or more files and/or folders to the recycle bin. Items can be restored later using the restore endpoint. Required scope (integrations): documents.write

delete_documents_from_recycle_bin

ChatGPT
Permanently delete documents from recycle bin. Delete single or multiple files and/or folders from recycle bin. > [!NOTE] > This action is irreversible. Once the documents are deleted from the recycle bin, they cannot be restored. Required scope (integrations): documents.write

delete_groups

ChatGPT
Delete groups. Delete groups in bulk from the specified project Required scope (integrations): users.write

delete_users

ChatGPT
Delete participants. Bulk-deletes one or more participants (users) from a project. Destructive – participants lose their access. Resolve users IDs/emails (via get_users) and ask a confirmation from a user to execute this.

download_file

ChatGPT
Download a file from a project. Returns the file content as base64 (max 50 MB). For larger files, use the Ideals web interface / Public API. To download a file using the Ideals UI, resolve a link to the file first using the get_link_templates tool. To read or analyze what a file says, use read_file_content instead — only call this tool when the user explicitly asks to download the file or needs the raw bytes.

export_activity_logs

ChatGPT
Export activity log. Export the project activity log (the audit trail of who did what, and when) as an Excel file. Supports the same filters as get_activity_logs.

export_document_index

ChatGPT
Export document index. To summarise a project, describe it, or give an overview of what it contains, call this directly — it is the single call that answers project-level content questions. You do not need Get project details first. It exports the project's full document index — every file and folder with names, sizes, upload dates, labels, direct links, and the complete folder tree, for all items the authenticated user can access — in one call. Use it to answer: - Summarise / overview of a project — what files and folders it holds and how it's organised. - How big is the project / how many GB — sum the per-item sizes. - How many files or folders — count the rows. - What is the folder structure / the whole tree — every folder and file in one call. - Spot gaps or empty folders — folders with no files beneath them. Use this — not Get project details, which returns only metadata (title, status, dates) with no contents, counts, or sizes.

get_activity_logs

ChatGPT
Get activity log. Get a detailed report of all actions (most recent first) performed in the project. Call this directly for any "who viewed / accessed / downloaded", "what happened", or "recent activity" question.

get_attributes

ChatGPT
Get list of project attributes. Get a list of project attributes defined for the given corporate account. Does not contain real values of the attributes assigned to the projects. Required scope (integrations): accounts.read

get_corporate_accounts

ChatGPT
Get list of corporate accounts. Returns corporate accounts that the authenticated user has administrative access to Required scope (integrations): accounts.read

get_current_user

ChatGPT
Get current user. Returns the identity and profile of the authenticated user, including name, email, company and job title. Use to confirm who you are acting as, or to resolve 'me'/'my' in other requests, e.g. "use my email" or "use my company".

get_document_summary

ChatGPT
Create document summary. Returns an AI-generated summary of a single document. If summarization is disabled in project (errorCode 1031), call read_file_content instead and summarize the text yourself.

get_folder_documents

ChatGPT
Get list of folder documents. Returns all direct children (files and folders) of the specified folder. No pagination. Useful when you only need the full list of direct children without traversing subfolders (top root-level folders by default). To get every file and folder across the entire project in one call, use export_document_index instead.

get_group_details

ChatGPT
Get group details. Get all details such as role, feature permissions, and security settings of the given group. Required scope (integrations): users.read

get_groups

ChatGPT
Get list of groups. Get a list of user groups that are present in the specified project. Required scope (integrations): users.read

get_label_groups

ChatGPT
Get list of label groups. Returns all label groups and their labels defined for the specified project. Use to understand the tagging scheme or to report which labels exist. Required scope (integrations): documents.read

get_permissions_by_document

ChatGPT
Get document permissions across groups. Returns the permissions assigned to each group for a specific file or folder. Each item in the response represents one group and its access level (None, FenceView, View, SaveEncrypted, SavePdf, SaveOriginal, Upload, or Manage). Required scope (integrations): documents.read

get_permissions_by_folder_and_group

ChatGPT
Get folder permissions for a group. Returns the permission level assigned to a specific group for a specific folder, along with the permissions of all child items (sub-folders and files) within it. The permissions field reflects the effective access level: None, FenceView, View, SaveEncrypted, SavePdf, SaveOriginal, Upload, or Manage. Required scope (integrations): documents.read

get_permissions_report

ChatGPT
Get permissions report. Export a full permissions report for the project as an Excel file. The report contains a matrix of all groups and their access levels across every folder and document in the project — useful for compliance audits and access reviews. Required scope (integrations): documents.read

get_project_details_by_id

ChatGPT
Get project details. Returns only a project's own metadata (e.g. project's root folder ID), not its contents. Do not call this to summarise, describe, or give an overview of a project, or to get its size, counts, or folder structure — it exposes none of those. For any of those, use export_document_index instead.

get_projects

ChatGPT
Get list of projects. Returns the projects available to the current user through this integration, each with its id (UUID), name, and status. This is the entry point for any project task: resolve a project name to the id other endpoints require by matching on the name and reusing the returned id — look the id up here rather than asking for it. If an expected project is missing, the integration most likely is not enabled for it or current user does not have access to it.

get_qna_configuration

ChatGPT
Get Q&A settings. Returns the Q&A workflow configuration: step toggles, assigned users, categories, and question teams. Use to understand the Q&A setup before raising/answering questions or perform other operations with Q&A. Returns 404 Not Found with error code 3009 if the Q&A module is not enabled for this project.

get_qna_thread_details

ChatGPT
Get thread details. Read up to 50 Q&A threads in one request with every post the caller can see: questions, question drafts, final answers, internal notes, assignment notes, proposed answers, answers pending approval, rejected answers and rejection notes. Every post carries the files and folders attached to it that the caller can access. To read one thread, pass one ID. Recommended workflow: 1. Call search_qna_threads to find the threads. It returns only the question and the final answer of each conversation; use its postCounts and updatedAt to decide which threads to read in full. 2. Call this endpoint with their IDs. Which posts the caller sees, by Q&A role: - Answer Coordinator, Answer Approver — every post of every conversation. - Expert — the posts of their current assignment cycle in the threads assigned to them. After being unassigned, they keep the posts they took part in. - Question Drafter, Question Submitter — questions, final answers and the question drafts of their own team: the unsent initial question of a thread in Draft status, or a follow-up drafted on an answered thread, which keeps the thread in Draft status. An administrator viewing the project as one of them gets the initial draft as a Question and does not get a follow-up draft, as in the web application. Posts are a flat list per conversation, ordered by creation date: filter by type. text is plain text with formatting and line breaks removed. A person with id 0 is hidden by the project's anonymity settings and must not be shown as a real user. threads is ordered by thread creation date, not by the order of the request: match threads by id. IDs that do not exist, were deleted, or are not visible to the caller are listed in notFound; the request never fails because of such IDs, even when it contains a single ID. Returns 400 Bad Request when threadIds is empty, has more than 50 IDs, or contains an empty ID. Returns 403 Forbidden when the caller has no Q&A role in the project. Returns 404 Not Found with error code 3009 if the Q&A module is not enabled for this project.

get_qna_threads

ChatGPT
Get list of threads. Lists Q&A threads (questions) with pagination and sorting (default: most recently updated). Use for a quick overview of outstanding questions. For filtered queries (status, priority, category, keyword, dates) use search_qna_threads.

get_recycle_bin_documents

ChatGPT
Get documents from recycle bin. Returns all files and folders currently in the recycle bin for the specified project Required scope (integrations): documents.read

get_report_actions

ChatGPT
Get activity log action types. Returns a static catalog of action-type identifiers, grouped by category — reference data only. It contains no events and does not tell you what happened in a project or who did anything. Rarely changed: get_activity_logs and export_activity_logs answer activity questions on their own. Use this only to map a specific action name to its exact identifier when you need to restrict one of those endpoints to a particular action type via the actionTypes filter. Required scope (integrations): reports.read

get_terms_of_use

ChatGPT
Display a project's Terms of Use for the user to review and Accept. Call this when the user wants to review or accept a project's Terms of Use, or after another tool reports that acceptance is required. Renders the terms with an Accept button that records the user's acceptance. This tool only DISPLAYS the terms (read-only); the user accepts by clicking the button in the card, so it is safe to call without asking the user for permission first.

get_terms_of_use_content

ChatGPT
Internal helper for the Terms of Use widget. Fetches the agreement's HTML body. Never call this directly — use get_terms_of_use instead, which returns everything needed to display and accept the terms.

get_user_details

ChatGPT
Get user details. Returns a participant's full details – name, group, and project access. REQUIRED first step before Update user settings (full replacement). Also useful for access reviews.

get_users

ChatGPT
Get list of participants. Lists all participants in a project with their group. Use to resolve user IDs/emails/name before updates, deletions, and to get an overview of the access structure.

move_documents

ChatGPT
Move documents. Move one or more files and/or folders to a different folder within the same project. The originals no longer exist at their previous location. > [!NOTE] > Items are processed sequentially. If an item fails (for example, due to insufficient permissions), processing stops and the response contains only the items that succeeded before the failure. Required scope (integrations): documents.write

patch_group

ChatGPT
Partially update group settings. Update only specific properties of a given group Required scope (integrations): users.write

read_file_content

ChatGPT
Get the text content of a file in a project (max 50 MB of text). Returns the text as fileContent. Prefer this tool over download_file to read or analyze what a file says. Use download_file only when the user needs the original file or its raw bytes. If the error says DocumentHasNotConvertedYet, the text of the file is not ready yet. Do not retry the call; tell the user to try again later.

rename_documents

ChatGPT
Rename documents. Renames 2 or more files and/or folders in a project in a single call. For a single file use rename_file; for a single folder use rename_folder. Input: - id: identifier of the existing file or folder to rename - isFolder: true if the item is a folder, false if it is a file - index: current index of the item within its parent folder (e.g. "1.2.3") - newName: the desired new name Response: Returns a summary with the total count of successfully renamed and failed items, plus per-item details indicating success or failure reason for each entry. Notes: - Between 2 and 3000 items must be specified; requests outside that range are rejected. - An item with an invalid name is reported as a failed item rather than failing the whole request; other valid items are still renamed.

rename_file

ChatGPT
Partially update file metadata. Update only specific properties of the specified file Required scope (integrations): documents.write

rename_folder

ChatGPT
Partially update folder metadata. Update only specific properties of a given folder. Rename silently returns the old name if the target name already exists — verify the response. Required scope (integrations): documents.write

restore_documents_from_recycle_bin

ChatGPT
Restore documents from recycle bin. Restore single or multiple files and/or folders from recycle bin. Required scope (integrations): documents.write

search_documents

ChatGPT
Search documents. Recursive search for documents under a folder (defaults to root) by query and/or filters. With an empty/* query and no filter, returns only the direct children of the folder. Use to find files by name or content across the whole project folder tree, or as the lookup step before summarising or answering. For a complete structural inventory (sizes, full tree, gaps) use export_document_index instead.

search_qna_threads

ChatGPT
Search threads. Search Q&A threads with advanced filtering, sorting, and pagination. Supports filtering by status, priority, category, thread type, author, assigned expert, keyword search, and date ranges. Default page size is 30. Default sorting: UpdatedAt descending (most recently updated first). Sorting behavior: - When searchQuery is provided and sortBy is not specified, results are sorted by relevance. - When searchQuery is not provided, results are sorted by UpdatedAt descending by default. - An explicit sortBy value always takes precedence regardless of searchQuery. Category filter: If the category filter is not specified, all threads are returned regardless of their category. When specified, use the filter field to control behavior: - Uncategorized — returns only threads without a category assigned. - Categorized — returns only threads that have any category assigned. - Specific — returns threads belonging to the categories listed in ids. Usage examples: Keyword search (sorted by relevance automatically): { "searchQuery": "due diligence deadline" } Keyword search with explicit sorting override: { "searchQuery": "deadline", "sortBy": "CreatedAt", "sortOrder": "Asc" } Filter by category: { "category": { "filter": "Specific", "ids": ["<guid1>", "<guid2>"] } } Filter uncategorized threads: { "category": { "filter": "Uncategorized" } } Returns 404 Not Found with error code 3009 if the Q&A module is not enabled for this project.

update_group

ChatGPT
Update group settings. Full replacement — all fields must be provided. Omitted fields are reset to their defaults. Always fetch the current group via GetGroupDetails first, then send all fields back with your changes applied. Required scope (integrations): users.write

update_user

ChatGPT
Update user settings. Full replacement of a single participant. All fields must be provided (group membership, document-notification frequency, etc.). Omitted fields reset to defaults so always fetch the user via get_user_details first and resend all fields with your change applied. Required scope (integrations): users.write

add_users_to_project

Claude

answer_question

Claude

ask_question

Claude

cancel_closure_schedule

Claude

close_project

Claude

copy_documents

Claude

create_folder

Claude

create_folders

Claude

create_group

Claude

create_project

Claude

delete_documents

Claude

delete_documents_from_recycle_bin

Claude

delete_groups

Claude

delete_users

Claude

download_file

Claude

export_activity_logs

Claude

export_document_index

Claude

get_activity_logs

Claude

get_attributes

Claude

get_corporate_accounts

Claude

get_current_user

Claude

get_document_summary

Claude

get_folder_documents

Claude

get_group_details

Claude

get_groups

Claude

get_label_groups

Claude

get_permissions_by_document

Claude

get_permissions_by_folder_and_group

Claude

get_permissions_report

Claude

get_project_details_by_id

Claude

get_projects

Claude

get_qna_configuration

Claude

get_qna_threads

Claude

get_recycle_bin_documents

Claude

get_report_actions

Claude

get_terms_of_use

Claude

get_user_details

Claude

get_users

Claude

move_documents

Claude

patch_group

Claude

read_file_content

Claude

rename_documents

Claude

rename_file

Claude

rename_folder

Claude

restore_documents_from_recycle_bin

Claude

search_documents

Claude

search_qna_threads

Claude

update_group

Claude

update_user

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/vdr-ideals"><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/vdr-ideals)