Avleon Ai

Warning: Avleon does not ship an AI integration yet. There is no @avleon/core API for LLM providers at this time — this page is a placeholder for planned work.

In the meantime you can call any provider's SDK directly from a service, the same as any other outbound HTTP dependency:

import { AppService } from '@avleon/core';

@AppService
export class SummaryService {
  async summarize(text: string) {
    // call your provider's SDK here
    return { summary: text.slice(0, 120) };
  }
}

Long-running generations are a good fit for a queue, so the request does not block on the model, and events can notify clients over websockets when a result is ready.