Life Sciences data in mixed reality
An AI agent working inside an Alepou project queries Patsnap's Life Sciences MCP servers, builds what it needs on the PC (pages, a 3D pipeline chart, molecules computed from Patsnap structures, a patent family) and places it in the room. Nothing in the headset is specific to Patsnap or chemistry: it is a generic viewer for what the project sends.
One take in the headset, about six minutes, lightly edited for pace. There is no need to watch it before we speak.
0:04Introduction0:41Opening the Patsnap app0:56The story: Patsnap data on the obesity race2:26Who is winning the GLP-1 pill race?3:06"Show me orforglipron"3:34"Make it bigger"3:57What does the patent cover?5:00Just the shared core5:23One variant on the core
How it works
A spoken request in the headset reaches an AI agent working inside an Alepou project on the PC.
The agent calls Patsnap's MCP servers: Pharma Intelligence for drugs, deals and patents; Chemical Molecular for structures and ADMET predictions.
Project tools turn the answers into pages, charts and 3D molecules. Every call, file and change is recorded in the project.
Alepou checks each file and places it in the room. The headset simply shows it.
What is live and what is computed
- Drug, company, deal and patent facts come from Patsnap's MCP servers, retrieved on 3 and 4 October 2026.
- Molecular structures come from Patsnap Chemical Molecular. The 3D shapes are computed on the PC (RDKit), not measured or receptor-bound structures.
- ADMET values are Patsnap's model predictions: screening signals, not measurements.
- The patent family's shared core and the six variants shown were selected computationally from the structures Patsnap lists in the patent.
- To keep the recording short, the molecules were built earlier the same day with the same tools; a new compound takes about ten seconds.
- This is an independent demonstration. It is not affiliated with or endorsed by Patsnap.