Base Agent

Required Parameters

• Template: Message template
• Handle Parse Errors: Error handling
• Max Tokens: Token limit
• Model Name: Language model
• JSON Mode: JSON output mode
• OpenAI API Key: API authentication
• OpenAI API Base: API endpoint
• Temperature: Response randomness
• Seed: Random seed
• Max Retries: Retry attempts
• Agent Instructions: Behavior guidelines
• Web Search: Search capability
• Agent Description: Role description
• Language Model: Model configuration

Output Format

{
  "response": {
    "content": string,
    "metadata": {
      "model": string,
      "tokens": {
        "prompt": number,
        "completion": number,
        "total": number
      },
      "processingTime": string
    }
  }
}

Example Usage

const baseAgent = new BaseAgent({
  template: "{role}: {content}",
  handleParseErrors: true,
  maxTokens: 2048,
  modelName: "gpt-4",
  jsonMode: false,
  openaiApiKey: "YOUR_API_KEY",
  openaiApiBase: "https://api.openai.com/v1",
  temperature: 0.7,
  seed: 42,
  maxRetries: 3,
  agentInstructions: "You are a helpful assistant...",
  webSearch: {
    enabled: true,
    maxResults: 5
  },
  agentDescription: "A versatile AI assistant",
  languageModel: {
    type: "openai",
    version: "latest"
  }
});

const result = await baseAgent.process({
  input: "Tell me about artificial intelligence"
});