MCP SERVERSearch an Outline wiki, read only the section you need, edit documents by string replacement, and reorganise a whole shelf of them in one batch.
LABSCO SUMMARY
A table of contents with line numbers, a section fetched by heading, grep inside one document, and string-match edits applied server-side add up to something usable on a wiki whose pages are too long to move around whole. The permission model is the other half: read-only, delete-disabled and scope-aware tool lists are three separate switches, so an assistant that only answers questions and one that reorganises collections are the same install with different settings.
A full read-and-write surface over an Outline knowledge base — cloud or self-hosted. It goes further than list-and-read: documents can be navigated by heading, edited without holding the whole text, and moved or archived in batches.
An Outline account, cloud or self-hosted, and an API key from Settings → API Keys. OUTLINE_API_KEY carries it, OUTLINE_API_URL points at a self-hosted instance, and both can live in ~/.config/mcp-outline/.env instead. Python 3.10+ outside Docker; it starts as uvx mcp-outline, or from the ghcr.io/vortiago/mcp-outline image. Permissions are configurable at the server: OUTLINE_READ_ONLY=true leaves only search, read, export and comment-viewing tools, OUTLINE_DISABLE_DELETE=true keeps create and update but removes the delete tools, and OUTLINE_DYNAMIC_TOOL_LIST=true hides tools the caller's API key scopes and Outline role would not allow. In HTTP mode each user can send their own key in an x-outline-api-key header, which takes precedence over the environment variable. Note that Outline now ships an official MCP server of its own. MIT.
One command plus a key — uvx mcp-outline, then supply credentials
USE CASES
Search for the document, look at its heading structure, and pull only the section that covers the question.
Find the exact line with a grep-style search, then replace that string in place — the rest of the document never moves.
List documents by most recent change, then read the ones that matter and the comments left on them.
Read the collection's structure, check what links to the documents you are about to move, then move or archive them in one batch.
Create the collection, add the documents under it — several at once when the outline is already settled — and publish them.
List the attachments the document points at, then either resolve a download URL or pull the file content back directly.
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
add_commentAdds a comment to a document or replies to an existing comment.
Use this tool when you need to:
Args: document_id: The document to comment on text: The comment text (supports markdown) parent_comment_id: Optional parent comment ID (for replies)
Returns: Result message with the new comment ID
archive_documentArchives a document to remove it from active use while preserving it.
IMPORTANT: Archived documents are removed from collections but remain searchable in the system. They won't appear in normal collection views but can still be found through search or the archive list.
Use this tool when you need to:
Args: document_id: The document ID to archive
Returns: Result message confirming archival
ask_ai_about_documentsQueries document content using natural language questions.
Use this tool when you need to:
Args: question: The natural language question to ask collection_id: Optional collection to limit the search to document_id: Optional document to limit the search to
Returns: AI-generated answer based on document content with sources
batch_archive_documentsArchives multiple documents in a single batch operation.
This tool processes each document sequentially, continuing even if individual operations fail. Rate limiting is handled automatically by the Outline client.
IMPORTANT: Archived documents are removed from collections but remain searchable. They won't appear in normal collection views but can still be found through search or the archive list.
Use this tool when you need to:
Recommended batch size: 10-50 documents per operation
Args: document_ids: List of document IDs to archive
Returns: Summary of batch operation with success/failure details
batch_create_documentsCreates multiple documents in a single batch operation.
This tool processes each creation sequentially, continuing even if individual operations fail. Rate limiting is handled automatically.
Use this tool when you need to:
Note: For Mermaid diagrams, use ``mermaidjs (not ``mermaid) as the code fence language identifier for proper rendering.
Recommended batch size: 10-50 documents per operation
Args: documents: List of document specifications
Returns: Summary of batch operation with created document IDs and success/failure details
batch_delete_documentsDeletes multiple documents, moving them to trash or permanently.
This tool processes each document sequentially, continuing even if individual operations fail. Rate limiting is handled automatically.
IMPORTANT: When permanent=False (the default), documents are moved to trash and retained for 30 days. Setting permanent=True bypasses trash and immediately deletes documents without recovery option.
Use this tool when you need to:
Recommended batch size: 10-50 documents per operation
Args: document_ids: List of document IDs to delete permanent: If True, permanently deletes without recovery option
Returns: Summary of batch operation with success/failure details
batch_move_documentsMoves multiple documents to a different collection or parent.
This tool processes each document sequentially, continuing even if individual operations fail. Rate limiting is handled automatically.
IMPORTANT: When moving documents that have child documents, all children will move along with them, maintaining hierarchical structure. You must specify either collection_id or parent_document_id (or both).
Use this tool when you need to:
Recommended batch size: 10-50 documents per operation
Args: document_ids: List of document IDs to move collection_id: Target collection ID (optional) parent_document_id: Target parent document ID (optional)
Returns: Summary of batch operation with success/failure details
batch_update_documentsUpdates multiple documents with different changes.
This tool processes each update sequentially, continuing even if individual operations fail. Rate limiting is handled automatically.
Use this tool when you need to:
Note: For Mermaid diagrams, use ``mermaidjs (not ``mermaid) as the code fence language identifier for proper rendering.
Recommended batch size: 10-50 documents per operation
Args: updates: List of update specifications
Returns: Summary of batch operation with success/failure details
create_collectionCreates a new collection for organizing documents.
Use this tool when you need to:
Args: name: Name for the collection description: Optional description of the collection's purpose color: Optional hex color code for visual identification (e.g. #FF0000)
Returns: Result message with the new collection ID
create_documentCreates a new document in a specified collection.
Use this tool when you need to:
Note: For Mermaid diagrams, use ``mermaidjs (not ``mermaid) as the code fence language identifier for proper rendering.
Args: title: The document title collection_id: The collection ID to create in text: Optional markdown content for the document parent_document_id: Optional parent document ID for nesting publish: Whether to publish immediately (True) or save as draft (False) template: If True, creates the document as a template icon: Optional emoji character to use as the document icon (e.g. "📋", "🚀"). If None, no icon is set.
Returns: Result message with the new document ID
delete_collectionPermanently removes a collection and all its documents.
Use this tool when you need to:
WARNING: This action cannot be undone and will delete all documents within the collection.
Args: collection_id: The collection ID to delete
Returns: Result message confirming deletion
delete_documentMoves a document to trash or permanently deletes it.
IMPORTANT: When permanent=False (the default), documents are moved to trash and retained for 30 days before being permanently deleted. During this period, they can be restored using the restore_document tool. Setting permanent=True bypasses the trash and immediately deletes the document without any recovery option.
Use this tool when you need to:
Args: document_id: The document ID to delete permanent: If True, permanently deletes the document without recovery option
Returns: Result message confirming deletion
edit_documentEdits a document using string-match replacement.
Each edit finds a unique old_string in the document and replaces it with new_string. Edits are applied server-side — you never need to hold the full document in context.
IMPORTANT: Batch all edits for a document into a single call when possible to minimize API calls. Each old_string must uniquely match one location in the document. Include surrounding context if needed to disambiguate.
Edits are applied sequentially, so later edits can target text created by earlier edits in the same batch. If any edit fails, no changes are applied (all-or-nothing).
By default (save=True), changes are pushed to Outline immediately. Set save=False to stage changes locally for large rewrites spanning multiple calls, then pass save=True on the final call to push all accumulated changes.
Args: document_id: The document ID to edit edits: List of edits to apply save: If True (default), push changes to Outline immediately. If False, stage locally.
Returns: Summary of edits applied and save status
export_all_collectionsExports the entire workspace content to a downloadable file.
IMPORTANT: This tool starts an asynchronous export operation which may take time to complete, especially for large workspaces. The function returns information about the operation, including its status. When the operation is complete, the file can be downloaded or accessed via Outline's UI. The export includes all collections, documents, and their hierarchies in the specified format.
Use this tool when you need to:
Args: format: Export format ("outline-markdown", "json", or "html")
Returns: Information about the export operation and how to access the file
export_collectionExports all documents in a collection to a downloadable file.
IMPORTANT: This tool starts an asynchronous export operation which may take time to complete. The function returns information about the operation, including its status. When the operation is complete, the file can be downloaded or accessed via Outline's UI. The export preserves the document hierarchy and includes all document content and structure in the specified format.
Use this tool when you need to:
Args: collection_id: The collection ID to export format: Export format ("outline-markdown", "json", or "html")
Returns: Information about the export operation and how to access the file
export_documentExports a document as plain markdown text.
Use this tool when you need to:
Args: document_id: The document ID to export
Returns: Document content in markdown format without additional formatting
fetch_attachmentFetch attachment content and return it as base64.
Calls attachments.redirect, follows the redirect, and returns the raw file content encoded as base64. Useful for images and files that agents can process.
Use this tool when you need to:
Args: attachment_id: The attachment UUID
Returns: Multi-line string in this format (blank line after Content-Length / before Content-Base64): Content-Type: <mime-type> Content-Length: <bytes>
Content-Base64: <base64-encoded-data>
Note: For large files (e.g. multi-MB PDFs), the base64 output may hit token limits. Prefer get_attachment_url to obtain a download URL for large attachments, then fetch externally.
get_attachment_urlResolve an attachment ID to a downloadable URL.
Calls attachments.redirect and returns the final URL after the redirect. Allows clients/agents to fetch the file themselves.
Use this tool when you need to:
Args: attachment_id: The attachment UUID
Returns: The redirect URL (signed download URL)
get_collection_structureRetrieves the hierarchical document structure of a collection.
Use this tool when you need to:
Args: collection_id: The collection ID to examine
Returns: Formatted string showing the hierarchical structure of documents
get_commentRetrieves a specific comment by its ID.
Use this tool when you need to:
Args: comment_id: The comment ID to retrieve include_anchor_text: Whether to include the document text that the comment refers to
Returns: Formatted string with the comment content and metadata
get_document_backlinksFinds all documents that link to a specific document.
Use this tool when you need to:
Args: document_id: The document ID to find backlinks for
Returns: Formatted string listing all documents that link to the specified document
get_document_id_from_titleLocates a document ID by searching for its title.
IMPORTANT: This tool first checks for exact title matches (case-insensitive). If none are found, it returns the best partial match instead. This is useful when you're not sure of the exact title but need to reference a document in other operations. Results are more accurate when you provide more of the actual title in your query.
Use this tool when you need to:
Args: query: Title to search for (can be exact or partial) collection_id: Optional collection to limit the search to
Returns: Document ID if found, or best match information
get_document_tocReturns the heading structure of a document as a table of contents with line numbers.
Use this tool when you need to:
Line numbers in the output can be used with read_document(offset=...) or read_document_section(heading=...) to read specific parts.
Args: document_id: The document ID
Returns: Formatted table of contents with line numbers
list_archived_documentsDisplays all documents that have been archived.
Use this tool when you need to:
Returns: Formatted string containing list of archived documents
list_collectionsRetrieves and displays all available collections in the workspace.
Use this tool when you need to:
Args: limit: Maximum number of results to return offset: Number of results to skip (pagination)
Returns: Formatted string containing collection names, IDs, and descriptions
If the response contains as many collections as the limit, execute the tool again with an increased offset to check for more results.
list_document_attachmentsList attachment IDs referenced in a document.
Parses document content for attachment references (e.g. /api/attachments.redirect?id=<uuid>) and returns a list of attachment IDs with context snippets.
Use this tool when you need to:
Args: document_id: The document ID to scan
Returns: Formatted list of attachment IDs and context
list_document_commentsRetrieves comments on a specific document with pagination support.
IMPORTANT: By default, this returns up to 25 comments at a time. If there are more than 25 comments on the document, you'll need to make multiple calls with different offset values to get all comments. The response will indicate if there are more comments available.
Use this tool when you need to:
Args: document_id: The document ID to get comments from include_anchor_text: Whether to include the document text that comments refer to limit: Maximum number of comments to return (default: 25) offset: Number of comments to skip for pagination (default: 0)
Returns: Formatted string containing comments with author, date, and optional anchor text
list_recently_updated_documentsLists documents ordered by most recent change (newest first).
Answers questions like "what changed this week" without needing search keywords. Backed by Outline's search endpoint with an empty query, sorted by last-modified time.
IMPORTANT: date_filter is a coarse server-side window on each document's last-modified time. It accepts only "day", "week", "month", or "year" (no arbitrary timestamps) and defaults to "week". Results are ordered newest-changed first.
STATUS FILTER: By default this lists published documents only. Pass status_filter to include other states. Allowed values are "draft", "archived", and "published". Drafts only ever include those you are allowed to see.
PAGINATION: Returns up to limit documents (default 25). Use offset to page through older changes.
Use this tool when you need to:
Args: date_filter: Time window on last-modified time. One of "day", "week", "month", "year". Defaults to "week". collection_id: Optional collection to limit results to status_filter: Optional list of statuses to include. Allowed values are "draft", "archived", and "published". Defaults to published only. limit: Maximum number of documents to return (default: 25) offset: Number of documents to skip for pagination (default: 0)
Returns: Formatted string listing documents with their IDs and last-updated timestamps, newest first
list_trashDisplays all documents currently in the trash.
Use this tool when you need to:
Returns: Formatted string containing list of documents in trash
move_documentRelocates a document to a different collection or parent document.
IMPORTANT: When moving a document that has child documents (nested documents), all child documents will move along with it, maintaining their hierarchical structure. You must specify either collection_id or parent_document_id (or both).
Use this tool when you need to:
Args: document_id: The document ID to move collection_id: Target collection ID (if moving between collections) parent_document_id: Optional parent document ID (for nesting)
Returns: Result message confirming the move operation
read_documentRetrieves and displays document content, optionally paginated by line range.
Use this tool when you need to:
When offset or limit are set, output includes line numbers and a metadata header showing the range. When both are 0 (default), the full document is returned in the original format for backward compatibility.
Args: document_id: The document ID to retrieve offset: 0-based line number to start from; must be non-negative (default: 0) limit: Number of lines to return; 0 means all lines; must be non-negative (default: 0)
Returns: Formatted string containing the document title and content
read_document_sectionReads a specific section of a document identified by heading match.
Uses case-insensitive substring matching against headings. Returns the section content from the matched heading up to the next heading of the same or higher level, including all nested subsections.
Use this tool when you need to:
Args: document_id: The document ID heading: Case-insensitive substring to match against headings (e.g. "arch" matches "## Architecture")
Returns: Section content with line numbers
restore_documentRecovers a document from the trash back to active status.
Use this tool when you need to:
Args: document_id: The document ID to restore
Returns: Result message confirming restoration
search_document_contentSearches within a document for lines containing a text snippet (grep-style), returning matches with line numbers and surrounding context.
Use this tool when you need to:
Matching is case-insensitive per line. Line numbers are 0-based and valid as read_document offsets.
Args: document_id: The document ID to search query: Text snippet to find (case-insensitive) context_lines: Lines of context around each match (default: 2, must be non-negative)
Returns: Matching lines with line numbers and context
search_documentsSearches for documents using keywords or phrases across your knowledge base.
IMPORTANT: The search performs full-text search across all document content and titles. Results are ranked by relevance, with exact matches and title matches typically ranked higher. The search will return snippets of content (context) where the search terms appear in the document. You can limit the search to a specific collection by providing the collection_id.
PAGINATION: By default, returns up to 25 results at a time. If more results exist, use the 'offset' parameter to fetch additional pages. For example, use offset=25 to get results 26-50, offset=50 for 51-75, etc.
STATUS FILTER: By default, searches published documents only. Pass status_filter to include other document states. Allowed values are "draft", "archived", and "published".
Use this tool when you need to:
Args: query: Search terms (e.g., "vacation policy" or "project plan") collection_id: Optional collection to limit the search to limit: Maximum results to return (default: 25, max: 100) offset: Number of results to skip for pagination (default: 0) status_filter: Optional list of statuses to search. Allowed values are "draft", "archived", and "published". Defaults to ["published"] when omitted.
Returns: Formatted string containing search results with document titles, contexts, and pagination information
unarchive_documentRestores a previously archived document to active status.
Use this tool when you need to:
Args: document_id: The document ID to unarchive
Returns: Result message confirming restoration
update_collectionModifies an existing collection's properties.
Use this tool when you need to:
Args: collection_id: The collection ID to update name: Optional new name for the collection description: Optional new description color: Optional new hex color code (e.g. #FF0000)
Returns: Result message confirming update
update_documentModifies an existing document's title or content.
IMPORTANT: This tool replaces the document content rather than just adding to it. For partial edits (changing specific text), prefer the edit_document tool instead.
Use this tool when you need to:
Note: For Mermaid diagrams, use ``mermaidjs (not ``mermaid) as the code fence language identifier for proper rendering.
Args: document_id: The document ID to update title: New title (if None, keeps existing title) text: New content (if None, keeps existing) append: If True, adds text to end of document instead of replacing template: If True, converts to a template. If False, converts a template back to a regular document. icon: Optional emoji character to use as the document icon (e.g. "📋", "🚀"). If None, keeps existing icon. Pass an empty string to remove the icon.
Returns: Result message confirming update
BEFORE YOU INSTALL
38 tools when we connected. OUTLINE_READ_ONLY=true leaves only read tools; OUTLINE_DISABLE_DELETE and OUTLINE_DISABLE_AI_TOOLS remove the delete tools and ask_ai_about_documents. OUTLINE_DYNAMIC_TOOL_LIST hides tools per user based on Outline role and API-key scopes.
CONTEXT COST
Rated by the size of the 38 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
uvx mcp-outline{
"mcpServers": {
"mcp-outline": {
"command": "uvx",
"args": [
"mcp-outline"
],
"env": {
"OUTLINE_API_KEY": "<YOUR_API_KEY>",
"OUTLINE_API_URL": "<YOUR_OUTLINE_URL>"
}
}
}
}claude mcp add mcp-outline uvx mcp-outlineThis command passes no API key. Set OUTLINE_API_KEY in your shell profile (~/.bashrc or ~/.zshrc) — plus OUTLINE_API_URL for self-hosted Outline — and restart Claude Code; without the key every tool call returns an error with setup instructions.
/plugin marketplace add Vortiago/mcp-outline
/plugin install mcp-outline@mcp-outlineSet your Outline API key in your shell profile (~/.bashrc or ~/.zshrc):
export OUTLINE_API_KEY="<YOUR_OUTLINE_API_KEY>"For self-hosted Outline also set OUTLINE_API_URL, then restart Claude Code
MCP servers
anuragagrawal0430
HelicopterHelicopter
cosmix
ArianChen777