top of page
Background Image

Accessible AI for molecule generation & optimisation

DeepMirror App is our intuitive AI powered cloud app that assists medicinal chemists who want to find effective and safe drug compounds by reducing the number of DMTA cycles required to find the optimal compound and suggesting creative and potent novel molecules.

Ultimately accelerating preclinical drug discovery by up to 50% and saving companies millions. 

Get in touch to request access to the app

Biologics (DNA/RNA and peptides) support coming soon.

Interface of DeepMirror Chem webapp
Interface of DeepMirror Chem webapp

Accessible AI

Do not waste time and money on AI. Get started with drug property and potency prediction within a few hours and see where AI can add value in your drug discovery programme.

Enhance creativity

Use generative AI to instantly create novel compounds to enhance creativity and escape flat SAR to reduce programme failure rates

50% faster Hit-to-candidate

Get from hit to lead candidate up to 50% faster by smart compound prioritisation and save millions in unnecessary synthesis


Our technology: Small Data AI

Dataset icon
Icons for a variety of Machine Learning model
Graph icon
Filled-in dataset icon

Upload your experimental data through our user-friendly interface

The best performing model is automatically selected for a given dataset

Experimental endpoints
are predicted using
the best model

A large library of different machine learning models are trained on a user's experimental data

Case Studies

Small Molecules
Antibodies & Peptides

We simulated small molecule lead optimisation workflows both with and without DeepMirror Chem and estimate that we enable users to find the most optimal drug up to ~4x faster.

4x acceleration of small molecule lead optimisation

Thumbnail of graph showing 4x drug discovery acceleration using AI

In this case study, we use DeepMirror Chem to learn and predict PROTAC properties to accelerate PROTAC discovery

Predict PROTAC affinity and degradation

Thumbnail showing general PROTAC structure and principle of action

DeepMirror Bio learns which gRNA sequence and secondary structure motifs correlate with high prime editing efficiency. Predictions can then be used to select optimal guides.

Identifying optimal gRNAs for Prime Editing

DeepMirror uses structural prediction algorithms (such as AlphaFold) and Graph Machine Learning to learn protein properties such as the yield in a cell free protein expression system.

Coming Soon...

Predicting cell free protein expression yield

in csv and sdf format

Upload and merge datasets seamlessly

molecules based on molecular scaffolds

Visualise, align, and compare molecules

based on your dataset with generative AI

Generate new molecules

molecules to compare them

and prioritise

Filter, sort and pin

such as Ligand Efficiency and Synthetic Accessibility

Calculate molecular properties

using cutting-edge AI from as few as 40 compounds

Predict properties

molecules based on predictions and confidence estimates


with your team or collaborators

Share results

bottom of page