An-tho is an independent applied AI lab. Our work follows two directions:

1. Building AI applications people can trust, through careful AI engineering.
2. Learning at the frontier of AI by building, and sharing what we learn.

Our first project, Vela, was a multilingual, citation-first medical search tool for healthcare professionals who work beyond English. Every answer pointed to its sources, and anything the model judged on its own was labelled as such. Vela was archived in October 2026; a demo of its Research feature remains online.

We share research notes and observations on the AI industry on LinkedIn. For anything else, write to us at support@an-tho.com.

An-tho 是一個獨立的應用 AI 實驗室，我們專注於兩個方向：

1. 以紮實的 AI 工程，打造值得信任的應用。
2. 在 AI 前沿動手實作，並分享我們從中學到與知道的事。

我們的第一個專案 Vela，是為跨語言工作的醫療專業人員打造、以出處為先的多語言醫學搜尋工具。每個回答都附上來源，凡是模型自行判斷的內容都會清楚標示。Vela 已於 2026 年 10 月封存，「研究」功能的展示仍可使用。

我們的研究筆記與 AI 產業觀察發表在 LinkedIn。其他事情歡迎來信 support@an-tho.com。