Assistant API

The Intelena dapp ships an AI assistant grounded in the whitepaper. It is exposed by the dapp's Worker as a streaming endpoint.

POST /api/ai/chat

Request

{
  "messages": [
    { "role": "user", "content": "What is a Leak Score?" }
  ],
  "model": "opus"
}

• messages — alternating user / assistant turns. The last turn must be from the user.
• model — optional: opus (default, Claude Opus 5 with adaptive thinking), sonnet, or haiku.
• History is capped server-side (last 20 turns, 4,000 characters per message).

Response

200 text/plain — the reply is streamed as UTF-8 text chunks as they are generated. Read the body as a stream and append chunks to the UI.

Behaviour

• Answers are grounded in the whitepaper. Figures that are not yet published (token supply, audits, the $INTEL contract address) are reported as such rather than invented.
• The assistant never asks for private keys, seed phrases, viewing keys, or note data.
• It does not give investment advice.
• It never states or guesses a $INTEL contract address, and does not describe any fee, tax or buyback mechanism.

In the app, conversations, the active chat and the chosen model are kept in your browser only; replies render Markdown with copyable code blocks.

Client example

const res = await fetch("/api/ai/chat", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({ messages }),
});
const reader = res.body!.getReader();
const decoder = new TextDecoder();
for (;;) {
  const { value, done } = await reader.read();
  if (done) break;
  append(decoder.decode(value, { stream: true }));
}