Skip to main content

Documentation Index

Fetch the complete documentation index at: https://agent37.com/docs/llms.txt

Use this file to discover all available pages before exploring further.

Send stream: true on a message and the response comes back as Server-Sent Events. Each event is named, so you can render text, reasoning, and tool activity live. Events arrive in order, and the terminal response.completed event carries the final output_text and usage.

​
Events

EventPayload
response.created{ id, instance_id, session_id }
response.reasoning.delta{ text }, a chunk of the agent’s thinking
response.output_text.delta{ text }, a chunk of the visible answer
response.tool_call.started{ tool, label }
response.tool_call.completed{ tool, duration_ms }
response.tool_call.failed{ tool, error }, the run continues
response.completed{ output_text, usage }, terminal success
response.failed{ error: { code, message } }, terminal error
These events are the same whatever agent the instance runs (Hermes, Claude Code, Codex, OpenClaw). They are the streaming contract, not a per-agent detail, so your client code doesn’t change when you switch agents.
Reconnect safely. If a stream drops, reconnect to GET /v1/responses/{id}/stream. It replays a snapshot of everything so far, then resumes live, so the answer is never lost.

​
Parse the stream

Read the response body, split on the SSE frame boundary, and branch on each event’s event: line. Stop when you see a terminal event.
node
const res = await fetch("https://api.agent37.com/v1/responses", {
  method: "POST",
  headers: {
    Authorization: "Bearer sk_live_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    instance_id: "inst_x7",
    input: "Research the top 3 EV makers, write a memo.",
    stream: true,
  }),
});

const reader = res.body.getReader();
const decoder = new TextDecoder();
let buffer = "";

while (true) {
  const { value, done } = await reader.read();
  if (done) break;
  buffer += decoder.decode(value, { stream: true });

  // SSE frames are separated by a blank line
  const frames = buffer.split("\n\n");
  buffer = frames.pop(); // keep the trailing partial frame

  for (const frame of frames) {
    const event = frame.match(/^event: (.+)$/m)?.[1];
    const data = JSON.parse(frame.match(/^data: (.+)$/m)?.[1] ?? "{}");

    switch (event) {
      case "response.output_text.delta":
        process.stdout.write(data.text); // stream the answer
        break;
      case "response.reasoning.delta":
        // show the agent thinking, if you want
        break;
      case "response.tool_call.started":
        console.log(`\n[${data.tool}] ${data.label}`);
        break;
      case "response.completed":
        console.log("\nusage:", data.usage);
        break;
      case "response.failed":
        console.error("\nerror:", data.error);
        break;
    }
  }
}
Prefer not to stream? Send stream: false (the default) and the call returns the finished response as one JSON body, with the agent’s reply in output_text.

​
Next steps

Send a message

Every field on the core call, plus the response shape.

Build a chat app

Stream every reply into a per-user chat.