Skip to main content

​
TDAIMemory

Enterprise-grade cloud-based memory storage with TDAI services integration for production applications.

​
Import

import { TDAIMemory } from '@ag-kit/agents/storage';

​
Constructor

constructor(options: TDAIMemoryOptions & IBaseMemoryOptions)

​
Configuration

​
options
TDAIMemoryOptions & IBaseMemoryOptions
required
TDAI Memory configuration options extending BaseMemory options
Note: Either client or clientOptions must be provided.

​
Basic Usage

// Initialize with TDAI credentials
const memory = new TDAIMemory({
  sessionId: 'user-session-123',
  clientOptions: {
    apiKey: 'your-tdai-api-key',
    baseURL: 'https://api.tdai.tencent.com'
  },
  useCache: true,
  enableContextManagement: true
});

// Add and retrieve events
await memory.add({
  message: {
    id: 'msg-1',
    role: 'user',
    content: 'Hello, I need help with my project',
    timestamp: new Date()
  },
  state: { userId: 'user-123' }
});

const events = await memory.list({ limit: 10, maxTokens: 4000 });

​
Core Interface Methods

TDAIMemory inherits from BaseMemory and implements all standard memory interface methods (branching not supported). For complete API documentation including list(), add(), addList(), delete(), retrieve(), clear(), getCount(), isEmpty(), see the BaseMemory API Reference.

​
Caching

// Enable caching for better performance
const cachedMemory = new TDAIMemory({
  sessionId: 'session-123',
  clientOptions: { apiKey: 'your-key' },
  useCache: true  // Enables local caching
});

// Cache is automatically managed:
// - Populated on first read
// - Invalidated on write operations
// - Reduces API calls for repeated reads

​
Unique Features

  • Cloud Persistence: Automatic synchronization to TDAI cloud services
  • Cross-Device Access: Share sessions across multiple instances and devices
  • Enterprise Security: Built-in authentication and encryption
  • High Availability: Redundant cloud infrastructure
  • Local Caching: Optional performance optimization with cache invalidation
  • Production Scalability: Handles enterprise-scale deployments

​
Branching Limitations

TDAIMemory does not support session branching operations due to cloud service limitations.

​
Error Handling

try {
  await memory.add(event);
} catch (error) {
  if (error.code === 'TDAI_AUTH_ERROR') {
    // Handle authentication issues
    console.error('Invalid API key or expired token');
  } else if (error.code === 'TDAI_QUOTA_EXCEEDED') {
    // Handle quota limits
    console.error('API quota exceeded');
  } else if (error.code === 'TDAI_NETWORK_ERROR') {
    // Handle network issues
    console.error('Network connectivity problem');
  }
}

​
Limitations

  • No Session Branching: Branching operations are not supported
  • Network Dependency: Requires internet connectivity for all operations
  • API Quotas: Subject to TDAI service usage limits
  • Latency: Network calls introduce latency compared to local storage
  • Cost: Usage-based pricing for cloud storage and API calls