PharmaInformatic
A family of six AI expert systems that predict human and animal pharmacokinetics from a molecule's chemical structure alone, covering oral bioavailability, clearance, volume of distribution, half-life, absorption and plasma protein binding for preclinical drug candidates.
What is PharmaInformatic?
PharmaInformatic builds AI expert systems that predict how a drug candidate will behave in the body, starting from nothing more than its chemical structure. Six systems are published, each covering one pharmacokinetic parameter: IMPACT-C for clearance, IMPACT-D for volume of distribution, IMPACT-E for terminal elimination half-life, IMPACT-F for oral bioavailability, IMPACT-A for absorption and IMPACT-B for plasma protein binding. Depending on the system, predictions are produced for humans and for monkeys, dogs, rats and mice.
A distinction matters here, because the naming invites confusion. Each expert system rests on a matching knowledge base whose name drops the leading letter: PACT-C, PACT-D, PACT-E, PACT-F, PACT-A and PACT-B, where PACT stands for Preclinical And Clinical Trials Knowledge Base. The PACT databases are the curated data asset; the IMPACT systems are the products built on top of them. The publisher describes these databases as the world's largest structure-based pharmacokinetic collections, assembled over two decades by annotating preclinical and clinical results drawn from millions of scientific publications and covering, in its own words, a century of drug development knowledge up to 2025.
Together the six systems produce what the company calls a Full Pharmacokinetic Profile, or Full-PK: bioavailability in percent, half-life in hours, volume of distribution in litres per kilogram, clearance in millilitres per minute per kilogram and plasma protein binding in percent. No in vitro data, no synthesis and no animal study is required as input, which is the central claim of the offer and the reason it is presented as a replacement for animal testing rather than a complement to it.
Access is not self-service. There is no web application, no API, no documentation portal and no account to open. Work begins with a contact form, an evaluation request or a research collaboration agreement, and the publisher proposes that a prospective customer first validate the systems in-house against its own existing preclinical results, so that confidential structures never leave the company. Named collaborations with Avivia BV, UNIZYME Laboratories A/S and Bridge Bioresearch PLC illustrate the pattern, as do the publicly funded German research projects Full-PK-KI and Full-PK-KI-2.
What it does
- Predict human oral bioavailability from a candidate's chemical structure
- Predict clearance, volume of distribution and terminal elimination half-life
- Predict plasma protein binding and oral absorption of small molecules
- Rank structurally diverse candidates across therapeutic areas for prioritisation
- Estimate an optimal oral dose for a first-in-human clinical trial
- Replace pharmacokinetic experiments in rats, dogs and monkeys
- Support toxicity assessment and in vitro to in vivo extrapolation
When to use PharmaInformatic / When not to
A quick filter to help you decide if PharmaInformatic is the right fit.
When to use PharmaInformatic
- Medicinal chemists optimising small-molecule leads or prodrugs before committing to synthesis and animal work
- Preclinical DMPK and pharmacology teams that need a full ADME profile from structure only
- Biotech founders and R&D managers prioritising candidates across a portfolio with limited experimental budget
- Clinical development teams estimating a first-in-human oral dose or the odds of translational success
- Toxicology and safety groups needing plasma protein binding values for in vitro to in vivo extrapolation
When not to use PharmaInformatic
- Anyone expecting a self-service tool: there is no account, no web application and no API to call
- Teams working on biologics, peptides or other large molecules, since only small molecules are covered
- Buyers who need a published price before starting a conversation, as no rate card exists
- Clinicians or patients, because the systems address preclinical development rather than care
- Organisations requiring a certified vendor, since no ISO, SOC or GxP certification is claimed or evidenced
How to use PharmaInformatic
A typical end-to-end flow, from setup to results.
- Identify which pharmacokinetic parameter you need predicted and therefore which expert system applies
- Contact PharmaInformatic through the form on the contact page or by e-mail at contact_us@pharmainformatic.com
- Alternatively call +49 (0)4921-9933-60, or meet the team at an industry event such as BIO Europe Spring
- Describe your project, your therapeutic area and the candidates you want evaluated
- Agree the commercial terms, since no price or plan is published and each engagement is negotiated
- Run a feasibility check first: validate the system against preclinical results you already own for rats, dogs or monkeys
- Keep that validation in-house if confidentiality matters, so that structures stay within your company
- Once the accuracy is acceptable, submit the structures of the candidates you want profiled
- Use the returned Full-PK profile to rank candidates, optimise a prodrug or estimate a first-in-human dose
Pros & Cons
Pros
- Needs only a chemical structure, so candidates can be assessed before synthesis or any animal study
- Covers a complete pharmacokinetic profile rather than a single endpoint
- Cross-species predictions let a customer benchmark the systems against data it already owns
- Evaluation can be run in-house, so proprietary structures need not leave the company
- Backed by publicly funded German research programmes, including Full-PK-KI, Full-PK-KI-2 and EXITOX-II
- Recognised by the 2021 Herbert Stiller Award of 20,000 euros from Doctors Against Animal Experiments
- Three named customer collaborations, two of them with attributed quotes from the partner's chief executive
Cons
- No price, plan or billing information is published anywhere, and no pricing page exists
- No terms of service and no customer contract are published, only an imprint and a website privacy policy
- No self-service access at all: no web application, no API, no documentation and no mobile app
- No certification of any kind is claimed or evidenced, whether ISO, SOC or a pharmaceutical quality standard
- Accuracy claims are first-party and unquantified, with no independent benchmark published
- Small molecules only, so biologics and other large molecules fall outside the scope
- Site upkeep is visibly lagging: an unedited second footer lists products and pages that no longer exist, and two defaced pages remain live in the sitemap
Pricing & Plans
No price is published. There is no pricing page, no plan and no rate card anywhere on the site, and the complete sitemap contains no such document. Every call to action leads to a contact form, an evaluation request or a discussion about a research collaboration, so the commercial model is quotation on enquiry. No permanent free plan is announced, and no free trial is announced either: the evaluation and the in-house validation the site proposes are never described as free of charge. The only monetary figure on the site is the 20,000 euro Herbert Stiller research prize the founder received in 2021, which is an award and not a price.
Data, GDPR & hosting
A consolidated view of how PharmaInformatic handles your data.
GDPR overview
The privacy policy is a detailed German notice served in English and applies the GDPR explicitly. It cites Art. 6(1)(a), (b), (c) and (f) GDPR and Sect. 25(1) TDDDG as legal bases, and lists the full set of data subject rights: access, rectification, erasure, restriction, portability, objection under Art. 21 and complaint to a supervisory authority. Consent is revocable at any time and SSL/TLS encryption is stated. Hosting is with IONOS SE in Montabaur, Germany, under a signed data processing agreement, and a second agreement covers Google Analytics, whose transfers to the United States rely on Standard Contractual Clauses with IP anonymisation active. No data protection officer and no Art. 27 representative are named, which is consistent with a German controller. The site never claims in so many words to be GDPR compliant, and the policy covers the website only, not the prediction service.
Who owns the data?
No terms of service or customer contract is published, so data ownership for compounds submitted to the expert systems is nowhere defined. The only published document is a website privacy policy, and it covers visitors alone. It names Dr. Wolfgang Boomgaarden, PharmaInformatic Boomgaarden, Nesserlander Str. 92, 26723 Emden as the controller, reachable at contact_us@pharmainformatic.com. Product pages repeatedly promise that evaluations can be run so that no data or information leaves the customer's company, but that promise sits in marketing copy rather than in a contract. Anyone submitting proprietary structures should therefore negotiate confidentiality terms directly before sharing anything.
Reuse rights
Nothing on the site grants or restricts a right to reuse outputs, because no terms of service exist. For website data, the privacy policy states that information from the contact form and from e-mail, telephone or fax enquiries is used solely to handle the request and is not passed on without consent. Server log files are collected for technical operation and are not merged with other data sources. Google Analytics is used with consent to analyse browsing behaviour. What a customer may do with a predicted bioavailability or clearance value is left entirely to whatever agreement the two parties sign.
Data retention & training
Hosting summary
The only hosting disclosed anywhere concerns the website itself, not the prediction service. The privacy policy names IONOS SE, Elgendorfer Str. 57, 56410 Montabaur, Germany as the host, processing visitor log files under Art. 6(1)(f) GDPR and covered by a signed data processing agreement. Independent checks agree: the domain resolves to a German IONOS address in Frankfurt am Main and the server responds as a real origin rather than a content delivery node. One transfer leaves the European Union: Google Analytics sends website usage data to Google servers in the United States under Standard Contractual Clauses, with IP anonymisation switched on. Nothing at all is published about where customer compound structures or prediction results are processed or stored. The publisher instead states that a feasibility evaluation can be handled in-house so that no data leaves the customer's own company, which sidesteps the hosting question rather than answering it. Buyers with a data residency requirement should raise it explicitly before any engagement.
Things to keep in mind
Risks and trade-offs to weigh before adopting PharmaInformatic.
- Nothing contractual protects the structures you submit: the confidentiality promise lives in marketing copy, not in published terms, so secure it in writing first
- Accuracy claims are first-party and unquantified, so treat a prediction as a prioritisation signal and not as a substitute for the study a regulator will ask for
- Deciding to skip an animal study on a predicted value shifts real risk onto a model whose validation you have not seen
- The naming invites a costly mistake: PACT-F is a knowledge base, IMPACT-F is the product, and confusing the two misstates what you are buying
- The publisher appears to be a one-person business with no legal form or register number in its imprint, which is worth weighing for continuity on a multi-year programme
- Two defaced pages remain live in the site's own sitemap, a sign that the web presence is not actively monitored
- An unedited older footer advertises products that no longer exist, so verify the current catalogue directly rather than trusting the page you land on
Setup & Integrations
Technical difficulty
Technically there is nothing to set up. No software to install, no account to create, no key to configure and no integration to build: you submit a chemical structure and receive a prediction. The real effort is commercial and contractual, since every engagement starts with a conversation and is agreed case by case. The one preparatory step worth budgeting for is the in-house validation the publisher recommends, which means gathering your own preclinical results for rats, dogs or monkeys and comparing them against the system's output before you rely on it.
Behind PharmaInformatic
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What exactly do the PharmaInformatic systems predict?
What input do I have to provide?
What is the difference between IMPACT and PACT?
Which species are covered?
How much does it cost?
Is there a web application or an API?
Will my compound structures stay confidential?
Who is behind PharmaInformatic?
Is the technology publicly funded or independently recognised?
Which companies have already used the systems?
Should you pick PharmaInformatic?
PharmaInformatic is a narrow and deep tool rather than a general-purpose assistant. It answers one question well: what will this molecule do in the body, judged from its structure alone. For a medicinal chemistry or preclinical DMPK team, that question is expensive to answer any other way, and the six IMPACT systems cover the whole profile instead of a single endpoint.
The scientific credentials are unusually solid for a company of this size. Two consecutive German federal research programmes funded the work, the EXITOX-II consortium placed it alongside a university and a Fraunhofer institute, and a national award for animal-free research went to the founder in 2021. Three named customers describe concrete outcomes, two of them through their own chief executives.
Commercial transparency is the opposite. There is no price, no plan, no terms of service and no way to try anything without a conversation. Buyers who need to compare offers on published figures will find nothing to compare. The accuracy claims, including the assertion that prediction error is halved for protein binding, are the publisher's own and are not backed by a published benchmark. No certification is claimed, which for a supplier feeding decisions in regulated drug development is worth raising early.
Two housekeeping problems are visible from outside and deserve mention: an unedited older footer still lists products and pages that no longer exist, and two defaced pages remain live and indexed in the site's own sitemap. Neither affects the science, but both suggest the website is not closely watched.
Verdict: worth a conversation if you develop small molecules and can work under a collaboration agreement. Not worth your time if you expect to sign up and start today.
- Choosing a selection results in a full page refresh.
- Opens in a new window.