AppFlowy Cloud MCP Server

A Model Context Protocol (MCP) server for interacting with AppFlowy Cloud API, providing tools for workspace, database, and row operations.

• Authentication: Login and refresh token management
• Workspace Operations: List all workspaces
• Database Operations: List databases, get database fields
• Row Operations: List rows, get row details, create rows, upsert rows

The server uses in-memory token storage. To authenticate:

1. Use appflowy_login with your email and password
2. The tokens are stored automatically
3. Use appflowy_refresh_token when the access token expires

• appflowy_login(request: LoginRequest) - Login to AppFlowy Cloud
• appflowy_refresh_token(request: RefreshTokenRequest) - Refresh access token

• appflowy_list_workspaces() - List all workspaces
• appflowy_get_workspace_folder(workspace_id: str, depth: int | None = None, root_view_id: str | None = None) - Get workspace folder metadata

• appflowy_list_databases(workspace_id: str) - List databases in a workspace
• appflowy_get_database_fields(workspace_id: str, database_id: str) - Get database fields

• appflowy_list_rows(workspace_id: str, database_id: str) - List row IDs
• appflowy_get_row_details(workspace_id: str, database_id: str, row_ids: str, with_doc: bool = False) - Get row details
• appflowy_create_row(workspace_id: str, database_id: str, request: RowCreateRequest) - Create a new row
• appflowy_upsert_row(workspace_id: str, database_id: str, request: RowUpdateRequest) - Update or create row

• appflowy_create_collab(...)
• appflowy_update_collab(...)
• appflowy_get_collab(...)
• appflowy_get_collab_json(...)
• appflowy_batch_create_collab(...)
• appflowy_full_sync_collab(...)
• appflowy_web_update_collab(...)
• appflowy_create_page_view(...)
• appflowy_get_page_view(...)
• appflowy_append_block_to_page(...)
• appflowy_create_orphaned_view(...)
• appflowy_duplicate_page(...)
• appflowy_create_quick_note(...)
• appflowy_list_quick_notes(...)
• appflowy_search_documents(...)
• appflowy_publish_page(...)
• appflowy_unpublish_page(...)
• appflowy_import_zip(file_path: str)
• appflowy_create_import_task(...)

uv run python main.py

1. Login:

request = LoginRequest(email="your@example.com", password="your_password")
response = appflowy_login(request)

2. List workspaces:

workspaces = appflowy_list_workspaces()

3. Get database fields:

fields = appflowy_get_database_fields("workspace_id", "database_id")

4. Create a row:

row_request = RowCreateRequest(cells={"Field_Name": "Value"}, document="Optional markdown")
result = appflowy_create_row("workspace_id", "database_id", row_request)

The server maintains tokens in memory. For production use, consider adding persistent storage (Redis, database, etc.) and proper error handling.

if you are slef hosting and you want the mcp to work just added as env in where you get you AI