B4AI Knowledge Hub

Activepieces integration for managing B4AI Knowledge Hub - a graph-based knowledge management system with AI-powered schema generation.

Overview

The B4AI Knowledge Hub piece enables you to manage knowledge hubs, synchronize graph schemas, and leverage LLM-generated metadata for your graph databases. Each knowledge hub consists of graph-based data structures with nodes, relationships, and properties.

Features

Knowledge Hub Management

• List Knowledge Hubs - View all knowledge hubs in your workspace
• Get Knowledge Hub - Retrieve a specific hub with its complete schema
• Create Knowledge Hub - Create a new hub with DRAFT and PRODUCTION instances
• Update Knowledge Hub - Modify hub name or description
• Delete Knowledge Hub - Permanently remove a hub and all its data

Schema Management

• Sync Schema - Synchronize schema from Memgraph to PostgreSQL
• Get Schema JSON - Retrieve schema with LLM-generated metadata
• Regenerate Metadata - Trigger async regeneration of AI metadata

Authentication

This piece requires three authentication parameters:

1. Base URL: Your B4AI server URL (e.g., https://your-server.com)
2. API Key: Your B4AI API key (x-api-key header)
3. API Secret: Your B4AI API secret (x-api-secret header)

Getting Your Credentials

1. Log in to your B4AI platform
2. Navigate to Settings > API Keys
3. Generate a new API key pair
4. Copy both the API Key and API Secret
5. Note your server's base URL

Actions

CRUD Operations

List Knowledge Hubs

Get all knowledge hubs for your workspace.

Parameters: None

Returns: Array of knowledge hubs with id, name, description, and timestamps.

Get Knowledge Hub

Retrieve a specific knowledge hub with its complete schema including all node types, relationship types, and properties.

Parameters:

• hubId (required): UUID of the knowledge hub

Returns: Detailed knowledge hub object with schema structure.

Create Knowledge Hub

Create a new knowledge hub. This automatically creates both DRAFT and PRODUCTION Memgraph instances with a default "Entity" node type.

Parameters:

• name (required): Name of the knowledge hub
• description (optional): Description of the knowledge hub

Returns: Created knowledge hub object.

Update Knowledge Hub

Update the name or description of an existing knowledge hub.

Parameters:

• hubId (required): UUID of the knowledge hub
• name (optional): New name
• description (optional): New description

Returns: Updated knowledge hub object.

Delete Knowledge Hub

Permanently delete a knowledge hub and all associated data including schema, nodes, and relationships. This action is irreversible.

Parameters:

• hubId (required): UUID of the knowledge hub

Returns: Success message with warning about irreversibility.

Schema Management Operations

Sync Schema

Synchronize the schema from Memgraph to PostgreSQL. This captures the current state of node types, relationship types, and their properties.

Parameters:

• hubId (required): UUID of the knowledge hub
• generateMetadata (optional): Whether to generate LLM metadata during sync (default: false)

Returns: Success message and metadata generation status.

Get Schema JSON

Retrieve the knowledge hub schema with LLM-generated metadata including descriptions, search patterns, query examples, and common mistakes.

Parameters:

• hubId (required): UUID of the knowledge hub

Returns: Schema structure and metadata object with sync status.

Metadata Includes:

• Schema Text (detailed and simplified versions)
• Schema Description
• Search Patterns
• Plan Examples
• Query Examples
• Common Mistakes
• Sync Status (PENDING, SYNCING, SYNCED, ERROR, OUTDATED)
• Last Synced Timestamp

Regenerate Metadata

Trigger asynchronous regeneration of LLM metadata for the schema. This is an async operation.

Parameters:

• hubId (required): UUID of the knowledge hub
• mode (required): full or partial regeneration
• includeProduction (optional): Also regenerate for production instance (default: false)

Returns: Sync status and timestamp.

Note: This is an async operation. Use "Get Schema JSON" to check completion status.

Data Model

Knowledge Hub

{
  id: string;                    // UUID
  name: string;                  // Hub name
  description: string | null;    // Optional description
  workspaceId: string;           // Associated workspace
  createdAt: string;             // ISO timestamp
  updatedAt: string;             // ISO timestamp
}

Graph Schema Structure

Each knowledge hub contains:

• Node Types: Categories of entities (e.g., Person, Company, Product)
• Relationship Types: Connections between node types (e.g., WORKS_AT, OWNS)
• Properties: Attributes for nodes/relationships with data types

Property Data Types

• STRING - Text values
• INTEGER - Whole numbers
• FLOAT - Decimal numbers
• BOOLEAN - True/false values
• DATE - Date values
• DATETIME - Date and time values
• LIST - Array of values
• MAP - Key-value pairs

Usage Examples

Example 1: Create and Configure a Knowledge Hub

// Step 1: Create a new knowledge hub
const createResult = await createKnowledgeHub({
  name: "Company Knowledge Base",
  description: "Graph database for company entities and relationships"
});

const hubId = createResult.knowledgeHub.id;

// Step 2: Sync the schema
await syncSchema({
  hubId: hubId,
  generateMetadata: true
});

// Step 3: Get the schema with metadata
const schema = await getSchemaJson({
  hubId: hubId
});

Example 2: Check Schema Sync Status

// Get current schema and check sync status
const schemaData = await getSchemaJson({
  hubId: "your-hub-id"
});

console.log("Sync Status:", schemaData.metadata.syncStatus);
console.log("Last Synced:", schemaData.metadata.lastSyncedAt);

if (schemaData.metadata.syncStatus === 'OUTDATED') {
  // Regenerate metadata
  await regenerateMetadata({
    hubId: "your-hub-id",
    mode: "full",
    includeProduction: false
  });
}

Example 3: List and Manage Hubs

// List all knowledge hubs
const hubs = await listKnowledgeHubs();

console.log(`Found ${hubs.count} knowledge hubs`);

// Get details for each hub
for (const hub of hubs.knowledgeHubs) {
  const details = await getKnowledgeHub({
    hubId: hub.id
  });

  console.log(`Hub: ${details.knowledgeHub.name}`);
  console.log(`Node Types: ${details.knowledgeHub.nodeTypes.length}`);
  console.log(`Relationship Types: ${details.knowledgeHub.relationshipTypes.length}`);
}

Error Handling

The piece provides detailed error messages for common scenarios:

• 400 Bad Request: Invalid parameters (e.g., invalid UUID format)
• 401 Unauthorized: Invalid API credentials
• 403 Forbidden: Insufficient permissions
• 404 Not Found: Knowledge hub doesn't exist
• 409 Conflict: Knowledge hub name already exists
• 500 Server Error: Internal server error

Important Notes

1. Deletion is Permanent: Deleting a knowledge hub removes all data irreversibly
2. Async Operations: Metadata regeneration is asynchronous - poll with "Get Schema JSON"
3. Dual Instances: Each hub has DRAFT and PRODUCTION Memgraph instances
4. Schema Sync: Must sync schema after making changes in Memgraph
5. Metadata Generation: Optional but recommended for AI-powered features

Version

Current version: 1.0.0 (Complete revamp from document-based to graph-based knowledge management)

Support

For issues, questions, or feature requests, please contact the B4AI team or refer to the B4AI Knowledge Hub API documentation.