Introduction

The Vanguard AI API is a Computer Vision & AI platform built for Lao documents. It extracts structured data from images (OCR), transcribes and synthesises speech (STT / TTS), and exposes a multi‑model LLM chat — all behind a single API key.

Base URLs

Environment	Base URL
Production	https://ai-api.vanguardinitiative.com
Local development	http://localhost:8000

What you can do

• OCR — read passports, ID cards, driving licenses, bank cards and enterprise registrations into clean JSON.
• Speech‑to‑Text — transcribe Lao and other languages from audio.
• Text‑to‑Speech — synthesise natural speech from text.
• Images — create, edit, vary and upscale pictures from a prompt, in 12 styles and up to 4K.
• Chat — stream responses from multiple LLM providers with tool use.
• Agents — an agent you configure once that calls OCR, speech and search on its own.
• Responses — one model call with tools you define and run yourself, or JSON in a schema you set. The building block for an agent that works on your own data.
• Embeddings — turn text into vectors for semantic search and matching.

How the API is organised

• Control‑plane calls (auth, account, billing) use JSON.
• Document scans upload multipart/form-data.
• Long‑running work (STT, TTS) runs as asynchronous jobs you poll for completion.
• Every response uses the same envelope: { "status": "success" | "error", ... }.

Next steps

1. Follow the Quickstart to make your first call in minutes.
2. Learn how Authentication and API keys work.
3. Browse the API Reference for every endpoint.