Overview

Tinkuy supports streaming responses via agent.stream(). The agent sends tokens, tool call events, and completion events as they happen.

Usage

const agent = new Agent({
  router,
  tools: [myTool],
  systemPrompt: 'You are a helpful assistant.',
});

const stream = agent.stream('Tell me a story about AI');

for await (const chunk of stream) {
  switch (chunk.type) {
    case 'token':
      process.stdout.write(chunk.data);
      break;
    case 'tool_call':
      console.log('\n[Tool called:', chunk.data.tool, ']');
      break;
    case 'tool_result':
      console.log('\n[Tool result:', chunk.data, ']');
      break;
    case 'done':
      console.log('\n[Done]', chunk.data);
      break;
  }
}

Stream Events

Event	Data	Description
token	string	A text token from the LLM
tool_call	{ tool, args }	The LLM invoked a tool
tool_result	any	The tool’s execution result
done	AgentResult	Agent completed with final result

With React (frontend)

const response = await fetch('/api/agent', {
  method: 'POST',
  body: JSON.stringify({ prompt: 'Hello!' }),
});

const reader = response.body.getReader();
const decoder = new TextDecoder();

while (true) {
  const { done, value } = await reader.read();
  if (done) break;

  const text = decoder.decode(value);
  const lines = text.split('\n').filter(Boolean);

  for (const line of lines) {
    const chunk = JSON.parse(line.replace(/^data: /, ''));
    if (chunk.type === 'token') {
      // Append token to UI
    }
  }
}

On Complete Hook

const agent = new Agent({
  router,
  tools: [myTool],
  systemPrompt: '...',
  onComplete: (result) => {
    console.log('Agent finished:', {
      iterations: result.iterations,
      tools: result.toolsUsed,
      cost: result.costUsd,
      latency: result.latencyMs,
    });
  },
});