2b3pro roam research mcp

WHAT IT CONNECTS

2b3pro roam research mcp 为 AI Agent 提供什么

• Environment variable handling with .env support
• Comprehensive input validation
• Case-insensitive page title matching
• Recursive block reference resolution
• Markdown parsing and conversion
• Daily page integration
• Detailed debug logging
• Efficient batch operations
• Hierarchical outline creation
• `roam_fetch_page_by_title`: Fetch and read a page's content by title, recursively resolving block references up to 4 levels deep
• `roam_create_page`: Create new pages with optional content
• `roam_create_block`: Create new blocks in a page (defaults to today's daily page)
• `roam_import_markdown`: Import nested markdown content under specific blocks
• `roam_add_todo`: Add multiple todo items to today's daily page with checkbox syntax
• `roam_create_outline`: Create hierarchical outlines with proper nesting and structure
• `roam_search_block_refs`: Search for block references within pages or across the graph
• `roam_search_hierarchy`: Navigate and search through block parent-child relationships
• `roam_find_pages_modified_today`: Find all pages that have been modified since midnight today
• `roam_search_by_text`: Search for blocks containing specific text across all pages or within a specific page
• `roam_update_block`: Update block content with direct text or pattern-based transformations
• `roam_search_by_date`: Search for blocks and pages based on creation or modification dates
• `roam_search_for_tag`: Search for blocks containing specific tags with optional filtering by nearby tags
• `roam_remember`: Store and categorize memories or information with automatic tagging
• `roam_recall`: Recall memories of blocks marked with tag MEMORIES_TAG (see below) or blocks on page title of the same name
• `roam_datomic_query`: Execute custom Datalog queries on the Roam graph for advanced data retrieval and analysis
• Create a [Roam Research API token](https://x.com/RoamResearch/status/1789358175474327881):
• Go to your graph settings
• Navigate to the "API tokens" section (Settings > "Graph" tab > "API Tokens" section and click on the "+ New API Token" button)
• Create a new token
• Configure the environment variables:
• For Cline (`~/Library/Application Support/claude-code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json`):
• For Claude desktop app (`~/Library/Application Support/Claude/claude_desktop_config.json`):
• Build the server (make sure you're in the root directory of the MCP):
• Complete hierarchical structure
• Block references recursively resolved (up to 4 levels deep)
• Proper indentation for nesting levels
• Full markdown formatting
• `page_uid`: Direct reference to target page
• `title`: Name of target page (will be created if it doesn't exist)
• Neither: Block will be added to today's daily page
• Create complex outlines with up to 10 levels of nesting
• Validate outline structure and content
• Maintain proper parent-child relationships
• Optional header block for the outline
• Defaults to today's daily page if no page specified
• Efficient batch operations for creating blocks
• `outline`: Array of outline items, each with:
• `text`: Content of the outline item (required)
• `level`: Nesting level (1-10, required)
• `page_title_uid`: Target page title or UID (optional, defaults to today's page)
• `block_text_uid`: Header text for the outline (optional)
• Adds todos with Roam checkbox syntax (`{{TODO}} todo text`)
• Supports adding multiple todos in a single operation
• Uses batch actions for efficiency when adding >10 todos
• Automatically creates today's page if it doesn't exist
• Adds todos as top-level blocks in sequential order
• Import content under specific blocks:
• Find parent block by UID or exact string match
• Locate blocks within specific pages by title or UID
• Defaults to today's page if no page specified
• Control content placement:
• Add as first or last child of parent block
• Preserve hierarchical structure
• Efficient batch operations for nested content
• Comprehensive return value:
• `content`: Nested markdown content to import
• `page_uid`: UID of the page containing the parent block
• `page_title`: Title of the page containing the parent block (ignored if page_uid provided)
• `parent_uid`: UID of the parent block to add content under
• `parent_string`: Exact string content of the parent block (must provide either page_uid or page_title)
• `order`: Where to add the content ("first" or "last", defaults to "first")
• Find all references to a specific block
• Search for any block references within a page
• Search across the entire graph
• Supports both direct and indirect references
• Includes block content and location context
• `block_uid`: UID of the block to find references to (optional)
• `page_title_uid`: Title or UID of the page to search in (optional)
• Search for any text across all blocks in the graph
• Optional page-scoped search
• Case-sensitive or case-insensitive search
• Returns block content with page context
• Efficient text matching using Datalog queries
• `text`: The text to search for (required)
• `page_title_uid`: Title or UID of the page to search in (optional)
• `case_sensitive`: Whether to perform a case-sensitive search (optional, default: true to match Roam's native behavior)
• Two update modes:
• Direct content replacement
• Pattern-based transformation using regex
• Verify block existence before updating
• Return updated content in response
• Support for global or single-match replacements
• Preserve block relationships and metadata
• `block_uid`: UID of the block to update (required)
• `content`: New content for the block (if using direct replacement)
• `transform_pattern`: Pattern for transforming existing content:
• `find`: Text or regex pattern to find
• `replace`: Text to replace with
• `global`: Whether to replace all occurrences (default: true)
• Search for blocks containing specific tags
• Optional filtering by presence of another tag
• Page-scoped or graph-wide search
• Case-sensitive or case-insensitive search
• Returns block content with page context
• Efficient tag matching using Datalog queries
• `primary_tag`: The main tag to search for (required)
• `page_title_uid`: Title or UID of the page to search in (optional)
• `near_tag`: Another tag to filter results by (optional)
• `case_sensitive`: Whether to perform case-sensitive search (optional, default: true to match Roam's native behavior)
• Store information with #[[LLM/Memories]] tag
• Add optional category tags for organization
• Automatically adds to today's daily page
• Supports multiple categories per memory
• Easy retrieval using roam_search_for_tag
• Maintains chronological order of memories
• `memory`: The information to remember (required)
• `categories`: Optional array of categories to tag the memory with
• Search by creation date, modification date, or both
• Filter blocks, pages, or both
• Optional date range with start and end dates
• Include or exclude block/page content in results
• Sort results by timestamp
• Efficient date-based filtering using Datalog queries
• `start_date`: Start date in ISO format (YYYY-MM-DD) (required)
• `end_date`: End date in ISO format (YYYY-MM-DD) (optional)
• `type`: Whether to search by 'created', 'modified', or 'both' (required)
• `scope`: Whether to search 'blocks', 'pages', or 'both' (required)
• `include_content`: Whether to include the content of matching blocks/pages (optional, default: true)
• Tracks all modifications made to pages since midnight
• Detects changes at any level in the block hierarchy
• Returns unique list of modified page titles
• Includes count of modified pages
• No parameters required
• Direct access to Roam's query engine
• Support for all Datalog query features:
• Complex pattern matching
• Aggregation functions (count, sum, max, min, avg, distinct)
• String operations (includes?, starts-with?, ends-with?)
• Logical operations (<, >, <=, >=, =, not=)
• Rules for recursive queries
• Case-sensitive and case-insensitive search capabilities
• Efficient querying across the entire graph
• `query`: The Datalog query to execute (required)
• `inputs`: Optional array of input parameters for the query
• Count all pages:
• Case-insensitive text search:
• Find blocks modified after a date:
• Search up or down the block hierarchy
• Find children of a specific block
• Find parents of a specific block
• Configure search depth (1-10 levels)
• Optional page scope filtering
• Includes depth information for each result
• `parent_uid`: UID of the block to find children of (required if searching down)
• `child_uid`: UID of the block to find parents of (required if searching up)
• `page_title_uid`: Title or UID of the page to search in (optional)
• `max_depth`: How many levels deep to search (optional, default: 1, max: 10)
• Configuration errors:
• Missing API token or graph name
• Invalid environment variables
• API errors:
• Authentication failures
• Invalid requests
• Failed operations
• Tool-specific errors:
• Page not found (with case-insensitive search)
• Block not found by string match
• Invalid markdown format
• Missing required parameters
• Invalid outline structure or content
• Standard MCP error code
• Detailed error message
• Suggestions for resolution when applicable
• Install all required dependencies
• Compile TypeScript to JavaScript
• Make the output file executable
• Start the server in inspector mode
• Provide an interactive interface to:
• List available tools and resources
• Execute tools with custom parameters
• View tool responses and error handling