Reaction condition intelligence · starting with Suzuki–Miyaura

Suzuki conditions ranked by whether they reproduce, not just whether they were published once.

ReproRank mines pharmaceutical patents and shows you the protocols that independent organisations used on substrates like yours, with the evidence behind each one.

Built by a medicinal chemist, for medicinal chemists. Now in pilot.

live search · October 2026
Query · 4-bromoanisole + phenylboronic acid
1
Pd(dppf)Cl₂ · K₂CO₃ · dioxane/H₂O · 90 °C
6 organisations on similar substrates · precedent yield 100% · 132 applicants corpus-wide
REPRODUCED
2
Pd/PAd₃ · K₃PO₄ · THF/toluene · 25 °C
1 organisation on similar substrates · precedent yield 81% · close analogs
SINGLE SOURCE
3
Pd-PEPPSI/NHC · K₂CO₃ · EtOH · 25 °C
1 organisation on similar substrates · close analogs
SINGLE SOURCE
Every card links to the source patent and draws the precedent's structures.
How it works

One precedent is an anecdote. Several independent ones are evidence.

A patent tells you what worked once, for one team, on one substrate. When unrelated companies converge on the same catalyst, base and solvent for similar substrates, that protocol has been tested by many hands. ReproRank counts that convergence for you.

01 · QUERY

Enter your coupling

Paste your electrophile and boron partner, or the product. ReproRank finds structurally similar couplings in the patent literature.

02 · RANK

Count the evidence

Only complete procedures are shown and ranked by how many independent organisations used them on substrates like yours.

03 · RUN

Go to the bench

Instead of a list of 200 hits to sort through, you now leave with the best-supported protocol to try first and a backup or two. Set it up and go.

The dataset

Curated by a chemist, not scraped by a crawler.

19,587
Curated reactions
2,165
Pharma patents
2003–25
Coverage
REPRODUCED4 or more independent organisations ran this protocol on substrates like yours.
CONFIRMED2 or 3 organisations did.
SINGLE SOURCEOne organisation did. Real precedent, not yet confirmed by anyone else.
DISTANT PRECEDENTNobody ran it on substrates close to yours. Shown as a lead, never as consensus.
Matthieu Desroses
Matthieu Desroses
Founder · Medicinal chemist
Stockholm, Sweden
About

I have spent over twenty-five years in medicinal chemistry, and a painful share of that time repeating literature conditions that worked beautifully on paper and refused to work at the bench.

ReproRank started as my own answer to a simple question: of everything published, what has been reproduced? Every grouping decision in the dataset, what counts as the same chemistry, was made by hand.

Pilot

Try ReproRank on couplings you know well.

The pilot is free. You get a personal link and can run as many searches as you like for 30 days. Published or patented chemistry is ideal; nothing confidential is needed.

Request pilot access · pilot@reprorank.com

Structures and names you enter are never stored, not even as a hash. Hosted in the EU.
Already a pilot? Open the personal link you received again to get back in.