Skip to main content

​
Define an agent

Create a file in agents/: 
// agents/assistant.ts
import { agent } from "veryfront/agent";

export default agent({
  id: "assistant",
  model: "openai/gpt-4o",
  system: "You are a helpful assistant. Answer concisely.",
});
The id is how you reference the agent later with getAgent("assistant").

​
Add tools

Agents call tools to take actions or fetch data. Reference tools by name; the framework resolves them from the tools/ directory: 
// agents/assistant.ts
import { agent } from "veryfront/agent";

export default agent({
  id: "assistant",
  model: "openai/gpt-4o",
  system: "You are a weather assistant.",
  tools: { getWeather: true },
  maxSteps: 5,
});
maxSteps limits how many tool-call iterations the agent can perform per request. See Tools for how to define getWeather.

​
Connect to a route

Use getAgent() to retrieve a registered agent and stream its response: 
// app/api/chat/route.ts
import { getAgent } from "veryfront/agent";

export async function POST(request: Request) {
  const { messages } = await request.json();
  const agent = getAgent("assistant");
  const result = await agent.stream({ messages });
  return result.toDataStreamResponse();
}

​
Non-streaming response

For server-side generation (e.g., in getServerData), use generate(): 
import { getAgent } from "veryfront/agent";

const agent = getAgent("assistant");
const result = await agent.generate({
  input: "Summarize the latest news about AI.",
});

console.log(result.text);       // The agent's response
console.log(result.toolCalls);  // Tools the agent called
console.log(result.usage);      // Token usage

​
Dynamic system prompts

The system property accepts a string, a function, or an async function: 
export default agent({
  id: "assistant",
  model: "openai/gpt-4o",
  system: async () => {
    const date = new Date().toLocaleDateString();
    return `You are a helpful assistant. Today is ${date}.`;
  },
});

​
Agent configuration

PropertyTypeDescription
idstringUnique identifier used with getAgent()
modelstringProvider and model (e.g. "openai/gpt-4o", "anthropic/claude-sonnet-4-5-20250929")
systemstring | () => string | Promise<string>System prompt
toolsRecord<string, boolean | Tool>Tools the agent can use
maxStepsnumberMax tool-call iterations per request
memoryMemoryConfigConversation memory settings
streamingbooleanEnable streaming (default: true)
middlewareAgentMiddleware[]Execution middleware

​
Next