Platform

Search the database yourself. No account, no card.

The same data the paid analysis is built on, open to academics, scientists, industry teams and investors. Six ways to search, nothing behind a signup wall.

The platform sleeps when idle to avoid running a server for nobody, so the first request after a quiet period can take a few seconds to wake.

What is in it

Six ways to search, one catalogue.

Each search answers a different question about the same molecules, and each shows its working and its sources rather than returning a number alone. Where sources disagree in strength, the stronger one comes first.

Molecule

Look a drug up by name, ChEMBL identifier or SMILES and see what it treats, what it causes and what it resembles.

  • Labelled indications kept distinct from off-label and pipeline use
  • Adverse effects led by the literature: PubMed-indexed effects with PMIDs, then the US label, then spontaneous reports
  • Lipinski and physicochemical descriptors computed with RDKit
  • Tanimoto analogues over 2048-bit Morgan fingerprints

Adverse effect

Start from a reaction and see which drugs the literature links to it.

  • Plain clinical terms mapped to MeSH headings
  • Named drugs and drug classes, each with citable PMIDs
  • Most-cited Scopus papers alongside
  • Report counts shown last, as counts, never as incidence

Indication

What is used, or being investigated, for a condition.

  • Approved use separated from off-label and pipeline use
  • Clinical phase and modality on every hit

Interactions

The drug interaction section of the US label, retrieved live.

  • Single-ingredient label preferred over combination products
  • Label prose, not a pairwise checker, and labelled as such

Repurposing

Compounds with verified activity in more than one therapeutic domain, filtered to candidates worth a real look.

  • Cross-domain activity filtering across eight domains
  • Starting points for a repurposing hypothesis, not conclusions

Clinical agent

Ask a clinical pharmacology question in plain language and get a structured answer with resolvable citations.

  • Live PubMed retrieval runs before the answer is written
  • A synthesis of retrieved literature, not a memorised answer

What it runs on

Python, a Gradio interface over a SQLite catalogue of about 17,721 entities, RDKit for all chemistry, NCBI E-utilities for literature, and a retrieval-grounded language model for the clinical agent. The full method, including the docking, simulation and machine-learning work that sits behind the paid analysis, is written out on the About page.

What it is not

  • Not a clinical decision tool. Everything here is for research and development use. It is not medical advice and not a basis for treating a patient.
  • Not a regulatory database. Contents are a curated snapshot with a retrieval date, not an official source of record.
  • Not a replacement for interpretation. The platform will show you the signal; what it means for your compound, your protocol or your investment decision is the part that needs a person.

Found something in the platform worth a closer look? Send the compound or the trial number and the first response is a free one-page snapshot — request one here.

Tell us the compound and the question.

Scope and turnaround come back within two working days, with a personalised offer where the work needs a person. If the question is outside what the method can honestly answer, we will say so rather than sell you a report.