Managing agents via API

Every agent can also be driven programmatically. The API is OpenAI-compatible, so you can call it with a plain HTTP request or the standard OpenAI SDKs.

Endpoint & auth

• POST to your agent's endpoint with the path /api/v1/chat/completions
• Authenticate with a Bearer access key, scoped to that agent
• Send an OpenAI-style messages array; read the reply from choices[0].message.content

Example — cURL

curl -X POST "$AGENT_ENDPOINT/api/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer $AGENT_ACCESS_KEY" -d '{"messages":[{"role":"user","content":"Hello!"}]}'

Example — Python

import os, requests

resp = requests.post(
    os.environ["AGENT_ENDPOINT"] + "/api/v1/chat/completions",
    headers={"Authorization": "Bearer " + os.environ["AGENT_ACCESS_KEY"]},
    json={"messages": [{"role": "user", "content": "Hello!"}]},
)
print(resp.json()["choices"][0]["message"]["content"])

Example — JavaScript

const res = await fetch(process.env.AGENT_ENDPOINT + "/api/v1/chat/completions", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    Authorization: "Bearer " + process.env.AGENT_ACCESS_KEY,
  },
  body: JSON.stringify({ messages: [{ role: "user", content: "Hello!" }] }),
});
const data = await res.json();
console.log(data.choices[0].message.content);

Good to know

• Keep your access key server-side — never expose it in the browser
• Put per-IP or per-user rate limits in front of the API
• Streaming is supported for token-by-token responses