About

# Data infrastructure for medical AI

Mdexa turns scattered biomedical data into one database that can be searched, verified and called directly by agents. Every verdict is anchored in source text.

Every piece of data behind medical AI should have a clear source and a traceable record.

## How we work

-

### Evidence before verdicts

Each judgment carries the source text, its location and a verification date, so anyone can check it.

-

### Restrictions read broadly, permissions narrowly

We would rather flag one risk too many than let a hidden one pass.

-

### Built for agents

Data is not only for people to read; machines must be able to read, cite and check it directly.

[Methods, evaluations and notes from building the database: Research](https://mdexa.com/research/)

[Database](https://mdexa.com/database/) [Methodology](https://mdexa.com/methodology/) [Roadmap](https://mdexa.com/roadmap/) [GitHub](https://github.com/mdexa-ai)

- **Operator**: UNILINK AI PTE. LTD.

- **Registered in**: Singapore

- **Company registration no.**: 202610380N

- **Contact**: [support@mdexa.com](mailto:support@mdexa.com) or [GitHub issues](https://github.com/mdexa-ai/issues)