Fusebase MCP Server

Refresh Fusebase authentication cookies by re-launching a Playwright browser session. Use this when other tools return 401/auth errors. Set interactive=true to open a visible browser window for manual login. Returns cookie age and expiry info.

List all workspaces in your Fusebase organization with their IDs, titles, and colors. Use this first to discover workspace IDs needed by most other tools.

List pages (notes) in a Fusebase workspace with titles, types, and metadata. Supports pagination and filtering by folder. Returns compact summaries — use get_page for full details on a specific page.

Get detailed metadata for a specific page including title, dates, size, sharing status, and emoji. Returns all properties — use this when you need full page info beyond what list_pages provides.

Get recently accessed pages in a workspace, sorted by last access time. Useful for finding what the user was last working on. Returns up to the specified limit (default 10).

Create a new page in a Fusebase workspace with the given title. Optionally provide initial content via 'markdown' (recommended) or structured 'blocks'. Optionally specify a folderId to place it in a specific folder (defaults to root). Returns the created page's metadata including its new globalId.

List all folders in a Fusebase workspace as a nested tree structure. Each folder includes its children, icons, and sharing status. Use folder IDs to filter list_pages or as parentId when creating pages.

Get all attachments (images, files, audio recordings) embedded in a specific page. Returns file names, MIME types, sizes, and UUIDs. Useful for auditing media content or finding downloadable assets.

List all uploaded files across a workspace with names, sizes, formats, and URLs. Supports pagination with limit and offset. Different from get_page_attachments — this covers workspace-wide file storage.

Upload a file to a FuseBase page. The file content must be provided as base64-encoded data. Returns the attachment ID and URL path. Use get_page_attachments to list existing attachments.

Download a file attachment from a FuseBase page. Returns the file content as base64-encoded data with MIME type and size. Use get_page_attachments to find attachment IDs first.

Get all tags in a workspace, or tags for a specific page if pageId is provided. Workspace tags return the full tag vocabulary; page tags return only tags applied to that page. Use update_page_tags to modify a page's tags.

Set tags on a specific page, replacing any existing tags entirely. Pass an array of tag strings to apply. To remove all tags, pass an empty array.

Get members of a workspace or the entire organization if workspaceId is omitted. Workspace members include roles and privileges; org members include user profiles with emails. Useful for finding collaborators or checking permissions.

Get the HTML content of a page decoded from its Y.js document via WebSocket sync. Returns semantic HTML with headings, paragraphs, inline formats (bold, italic, code, links, strikethrough, underline), lists, blockquotes, code blocks, toggles, hints, collapsible headings, images, bookmarks, outlines,…

Search tasks in a workspace with full task details, assignees, labels, and board info. Optionally filter by page to see only tasks linked to a specific note. Supports pagination with offset and limit.

List task lists (kanban boards) in a workspace with their associated tasks, assignees, and reminders. Optionally filter to a specific task list by ID. Task list IDs are needed for create_task.

Create a new task in a workspace within a specified task list. Supports title, description, priority, due date, assignees, and labels. Use list_task_lists first to get valid taskListId values.

Get server version and check for updates from GitHub. Returns the current installed version, latest available version, whether an update is available, and the command to update. If update_available is true, inform the user that a new version is available and suggest updating.

Search the local FuseBase guide documentation (231 guides across 17 sections). Returns matching guide titles, sections, and slugs. Use get_guide to read the full content of a specific result. Great for looking up how any FuseBase feature works.

Get the full markdown content of a specific FuseBase guide by section and slug. Use search_guides first to find the right section/slug values. Returns the complete guide including instructions, screenshots, and hotkeys.

List all 17 FuseBase guide sections with the number of guides in each. Use this to browse available documentation categories before searching for specific topics.

Enable extended Fusebase tools for this session. By default only core tools (23) are loaded for performance. Call this with tier 'all' to dynamically register 68 additional tools for admin, analytics, content mutations, file upload, database CRUD, column management, and niche operations.

Get all labels (colored categories) defined in a workspace. Labels have titles, colors, and styles, and can be applied to tasks for visual organization. Returns label IDs usable in task creation.

Get organization-wide usage statistics including storage, traffic, member counts, AI credits, and workspace quotas. Each metric shows current vs max values. Useful for monitoring plan limits and resource consumption.

Get all comment threads on a specific page, including thread status (resolved/open) and nested comments. Useful for reviewing feedback, discussions, or collaborative annotations on a page.

Get the full rich-text description and detailed properties of a specific task. Use this when search_tasks provides insufficient detail. Returns the task's complete content including formatted description.

Create a new folder in a FuseBase workspace. Optionally specify a parentId to create a subfolder. Returns the created folder's metadata including its globalId.

Update a page or folder's properties — rename it, move it to a different folder, or both. Uses the upsert endpoint so partial updates are safe.

Update a task's properties — change status, priority, title, description, assignees, or due date. Uses PATCH semantics so only specified fields are changed.

Delete a task permanently from a workspace. This action is irreversible.

Delete a page permanently from a workspace. This action is irreversible — the page and its content will be lost. Use get_page first to verify you have the correct page before deleting.

Write or replace content on a page using the native Y.js WebSocket protocol. Accepts markdown (recommended) or structured content blocks. Supports: headings (H1/H2/H3), paragraphs, bold, italic, strikethrough, underline, inline code, links, highlight, bullet/numbered/checkbox lists, dividers, blockq…

List all AI agents configured in the organization with their titles, descriptions, and types. AI agents are custom assistants created in Fusebase's AI features. Returns agent IDs and metadata.

Get all mentionable entities (users, pages, folders) in a workspace for @-mention autocomplete. Includes member counts, workspace structure overview, and owner info. Useful for understanding workspace scope at a glance.

Get the full sidebar navigation tree showing all pages, folders, and their hierarchy across workspaces. Includes parent-child relationships, icons, and timestamps. Best way to get a complete structural overview of all content.

Get the activity feed for a workspace showing recent comments, @mentions, and content changes. Includes user avatars and note references for each activity item. Useful for monitoring workspace activity and collaboration.

Poll the Fusebase activity stream for new @mentions or comments directed at this profile. Use filterText to narrow results to only items mentioning a specific display name.

Create a new comment thread on a Fusebase page. The comment is anchored to a specific block (targetId) or to the page itself. Use this to leave feedback, ask questions, or communicate with human collaborators.

Reply to an existing comment thread on a Fusebase page. Use get_comment_threads first to find the thread ID, then reply to continue the conversation.

Resolve (close) a comment thread after it has been addressed. Use get_comment_threads to find thread IDs.

Get task usage statistics for a workspace, including upcoming deadline dates and active reminders. Useful for understanding task workload and scheduling pressure in a workspace.

Get recently updated notes across the entire organization, not limited to a single workspace. Returns notes sorted by last modification time with pagination support. Useful for finding the latest activity org-wide.

Get the total number of tasks in a workspace as a single count. Lightweight alternative to search_tasks when you only need the quantity, not the task details.

Get full workspace metadata including internal IDs, organization binding, creator userId, and creation/update timestamps. Provides deeper detail than list_workspaces. Useful for debugging or workspace administration.

Get the email-to-note addresses for a workspace. Sending emails to these addresses automatically creates pages in the workspace. Returns the dedicated email address and associated user.

Get the total count of files stored across all workspaces in the organization. Lightweight check for storage auditing — use list_files for detailed file listings.

Get AI feature usage for the organization showing current consumption vs maximum allowed. Tracks AI credits used across all workspaces. Useful for monitoring AI quota before heavy AI operations.

Get comprehensive organization permissions including all workspace memberships, role assignments, user avatars, and per-member usage data. More detailed than get_members — includes cross-workspace permission mapping.

Get workspace billing info including quota reset dates and plan details. Shows when usage counters reset and the organization's current billing cycle. Useful for understanding rate limits and renewal timing.

Get tags applied to a specific page/note. Unlike get_tags which returns workspace-wide tag vocabulary, this returns only the tags on one particular page. Use for checking a page's categorization.

Get structured data from a Fusebase database/table view. Requires the dashboard UUID and view UUID, which can be found in the URL when viewing a database in the Fusebase UI. Supports pagination via page and limit parameters.

List all databases/dashboards in the organization. Returns dashboard and view UUIDs that can be used with get_database_data. Probes known entity types (spaces, clients) plus any custom entities you specify.

Get data from a specific database entity type (e.g. clients, spaces). Automatically discovers the correct dashboard/view UUIDs for the entity. Supports pagination.

Create a new database (table or kanban view) in the organization. Returns the new database, dashboard, and view UUIDs. The database is created with a default table representation. Use list_databases afterwards to see it.

Add a new row to a database entity table. For built-in entities (clients, spaces), only entity name is needed. For custom databases, also provide databaseId and dashboardId (from list_databases or create_database). Uses Next.js server action internally.

Delete a row from a database. Requires the dashboard ID and the row ID. Use get_database_rows to find row IDs.

Move a kanban card to a different column by updating the grouped column's cell value. Equivalent to update_database_cell but semantically describes moving a card. Use get_database_schema to find the groupBy column key, and get_database_rows to find the row UUID.

List available relation targets and existing lookups for a dashboard. Returns linked tables/views that can be used for creating relation or lookup columns.

Create a new table (tab) within an existing database dashboard. This adds an additional table view alongside the existing one.

Delete a relation by its ID. This removes the link between two database tables. Use list_database_relations to find relation IDs.

List all databases in the organization via the dashboard-service REST API. Returns database metadata, dashboard (table) UUIDs, and view UUIDs. More comprehensive than list_databases — returns full database objects with nested dashboards.

Get detailed information about a specific database including all its dashboards (tables) and views. Use the database ID from list_all_databases or create_database.

Update a database's title, description, icon, color, favorite status, or public visibility. Uses PUT (PATCH is not supported by this API).

Delete a database and ALL its dashboards (tables), views, and data. This action is irreversible. Returns 204 on success.

Get detailed information about a dashboard (table within a database). Returns the dashboard's views array (including custom views), metadata, root entity, and scopes. Use the dashboard ID from list_all_databases or get_database_detail.

Delete a dashboard (table) within a database. Removes the table and its data. Use get_database_detail first to see available dashboards.

Update a custom view's name, filters, sorts, or hidden columns. Use get_dashboard_detail to find available views. Uses PUT (PATCH is not supported).

Switch a view's display mode. Supports 8 types: table (default spreadsheet), kanban (card columns), board, calendar, timeline, gallery, list, grid. Table and kanban use the representations endpoint; the rest use PUT with default_representation_template_id. Use get_dashboard_detail to find dashboard …

Duplicate (copy) an entire database, including tables, views, relations, and optionally data. Returns the new database with its UUIDs. Discovered via captured API: POST /databases/copy-from/database.

Create a new view within a dashboard (table). Each view can have its own filters, sorts, grouping, column visibility, and display mode. Use set_view_representation afterwards to change the view type.

Delete a view from a dashboard. Cannot delete the default (first) view. Use get_dashboard_detail to find view UUIDs.

Export a database view as CSV text. Returns the raw CSV content. Supports 5 delimiter options: comma, semicolon, pipe, tab, or caret. Use get_dashboard_detail to find dashboard and view UUIDs.

Duplicate an existing view within a dashboard. Creates a copy with the same schema, filters, and configuration. Use get_dashboard_detail to find dashboard and view UUIDs.

Import CSV data into an existing database table/view. Provide the CSV content, the database ID, dashboard (table) ID, and the view ID. The server will import the rows into the database. Column mapping is auto-generated from CSV headers (all imported as 'Single line text').

Set the grouping column for a kanban or board view. Groups cards/rows by the specified column. Requires the dashboard ID, view ID, and the column key to group by. Use get_database_schema first to find column keys.

Set the width of a column in a database view. Use get_database_schema first to find column keys. Width is in pixels.

Rename a column in a database view. Updates the column's name in the view schema. Use get_database_schema first to find the column key.

Reorder columns in a database view. Provide an array of column keys in the desired order. Columns not in the array are appended at the end. Use get_database_schema first to see current column keys and order.

Set the value of a specific cell in a database row. Use get_database_rows or get_database_data first to obtain the rowUuid and the column key (short opaque string like 'eoZSNDPy'). The get_database_rows tool returns a schema array mapping column names to keys for easy lookup. Note: rich-text (Descri…

Get rows from a database view in a structured format optimised for agent use. Each row includes: rowUuid (needed for update_database_cell), cells (column name → value map), and rawCells (column key → value map). Also returns the schema (column name, key, type) so you can look up the correct column k…

Get the column schema for a database view. Returns an array of column definitions with key (opaque 8-char ID, needed for update_database_cell), name (human-readable), type (string, number, date, label, checkbox, etc.), and edit settings. Use this to understand a database's structure before reading o…

Add a new column to a database view. Supported types: string (single-line text), multiline (multi-line text), number, date, label/status/select (with optional custom labels and colors), checkbox/boolean, email, phone, link/url, currency, files, user/assignee (assign org members), subtable/child-tabl…

Delete a column from a database view by its key. Use get_database_schema first to find the column key. This removes the column definition from the schema — existing cell data for that column key will no longer be visible. This action cannot be undone.

Add a Relation column that links this database to another database table. This creates a cross-database relation (via POST /relations) and adds a lookup-source column. The column shows linked records from the target table. Use get_database_schema on the target table first to get its dashboardId and …

Add a Lookup column that displays data from a related table through an existing Relation column. The lookup column is read-only and automatically pulls data from the linked records. You must have an existing relation column first (created via add_relation_column). Use get_database_schema to find the…

Get the organization's plan limits including maximum members, storage, traffic, AI credits, workspaces, and other quotas. Use alongside get_org_usage or get_usage_summary to compare current consumption against plan caps.

Get a condensed snapshot of organization usage vs limits — lighter than get_org_usage. Returns current/max for overall usage, storage, and blots in a single response.

List client portals for the organization. Optionally filter by workspace. Client portals are shared, branded pages published externally for clients or stakeholders.

Get pages published to a client portal. Returns the portal page tree for a given workspace, optionally filtered to a specific page. Use list_portals first to find active portals.

Get feature flags enabled for the organization. Each feature has an ID, name, and enabled status. Useful for checking what capabilities are available on the current plan.