ContextAgent

Agent with enhanced context management

ContextAgent

ContextAgent provides enhanced context management with RAG and conversation history.

Quick Start

import { createContextAgent } from 'praisonai';

const agent = createContextAgent({
  instructions: 'You are a helpful assistant.',
  llm: 'openai/gpt-4o-mini'
});

const response = await agent.chat('What is TypeScript?');
console.log(response.text);

Configuration

interface ContextAgentConfig {
  name?: string;
  instructions: string;
  llm?: string;
  knowledgeBase?: KnowledgeBase;
  contextWindow?: number;      // Max messages to keep
  maxContextTokens?: number;   // Max tokens in context
  verbose?: boolean;
}

With Knowledge Base

import { createContextAgent, createKnowledgeBase } from 'praisonai';

const kb = createKnowledgeBase({
  embeddings: 'openai/text-embedding-3-small'
});

await kb.add({ content: 'TypeScript is a typed superset of JavaScript.' });

const agent = createContextAgent({
  instructions: 'Answer questions using the knowledge base.',
  knowledgeBase: kb
});

const response = await agent.chat('What is TypeScript?');
// Response will include relevant context from knowledge base

Response Format

interface ContextResponse {
  text: string;
  context?: SearchResult[];  // RAG context if knowledge base is used
}

CLI Usage

praisonai-ts context chat "What is TypeScript?"
praisonai-ts context summarize "Long text to summarize..."
praisonai-ts context chat "Hello" --max-messages 20 --json