Drug Discovery AI
C

Certara

Certara sells biosimulation, pharmacometrics, chemical informatics and regulatory software to drug developers, with AI modules built into several product lines. Libra, launched in September 2025 inside the ToxStudio suite, predicts drug induced liver injury risk from a compound's structure and adds physicochemical properties and assay results as they arrive. CoAuthor drafts patient narratives, clinical study reports, protocols and other submission documents inside Microsoft Word from source files the writer selects, using a private GPT model for each customer along with eCTD templates and structured content. Certara.AI analyzes scientific documents and data, and AI PK Reports turns Phoenix tables, figures and listings into draft pharmacokinetic reports.

Writers can link source files straight from Veeva RIM through Certara's place in the Veeva AI Partner Program, and Chemaxon tools share chemical drawing and registration with D360. Certara.AI runs hosted or inside a customer's own infrastructure, while Libra is an Azure service.

Formed in 2008 from Tripos and Pharsight and based in Radnor, Pennsylvania, Certara counts more than 1,550 employees in over 30 countries, 94,000 users and 2,600 customer companies. In May 2026 it sold its regulatory and medical writing services business to Veristat, and CoAuthor remains a Certara product. Libra starts at $9,000 a year for one user; the other products are quoted.

AI Health Index verifiedNovember 19, 2026
Compare Certara with other vendors
Founded
2008
Headquarters
Radnor, Pennsylvania, United States
Website
www.certara.com
Categories
drug-discovery, clinical-trials-ai
Indexed Products
Libra, CoAuthor, Certara.AI, D360, Chemaxon, AI PK Reports
Buyer Segments
Pharmaceutical companies, Biotechnology companies, Contract research organizations, Regulatory writing teams
Assessment

Capability Axes

An AI Health Index grade measures what a buyer can verify from public sources on the date shown. It is not a rating of how good the product is. A vendor can build an excellent system and grade low on an axis because it publishes nothing an outsider can check. How grades read

AI Capability
BB on AI CentralityThe model is the engine of a core module. The platform carries other value, but this capability does not exist without it.
Vendor Published

Learned models sit inside specific products within a much larger modeling, software and consulting business. Libra pairs a QSAR model with a Bayesian machine learning model for early liver risk, CoAuthor generates regulatory drafts from approved source material, Certara.AI applies language models to scientific documents and data, and AI PK Reports drafts pharmacokinetic reports from Phoenix output. Those functions go without their models.

Simcyp physiologically based modeling, Phoenix pharmacometrics, Chemaxon drawing and registration, D360 data analysis and expert consulting keep their value without learned components, and mechanistic simulation is a different method from machine learning. The AI scope is the named predictive and generative modules rather than every calculation across the portfolio.

BB on Autonomy and Oversight ModelThe oversight structure is described and one part is missing, commonly the threshold at which the system stops or what happens after it is wrong.
Vendor Published

Medical writers pick CoAuthor's prompts and source documents, check the reference behind each generated passage and revise drafts in Word before anything moves toward submission. Collaborative review and version control keep that editing step in place, and the product is designed as human in the loop. Libra hands liver risk predictions to toxicology and discovery teams, who decide which compounds advance.

The Integral 7.1 MCP pilot lets external agents retrieve data only after an administrator enables the connection, with limits on write actions. No shared rule covers when a module withholds a low confidence prediction or routes a failed generation to a person, so review duties are clearer than behavior at those failure points.

BB on Model and Technology TransparencyThe approach or the suppliers are named without the version and update discipline behind them. Naming a supplier is the entry to this band both here and on Model Supply Chain Disclosure, which ask different questions of the same disclosure: who receives the data, and what produces the output.
Vendor Published

Libra names its methods: a QSAR model trained on 1,330 benchmark drugs and a Bayesian machine learning model that adds physicochemical and assay inputs, with the Max edition adding batch testing, data imputation and generated reports. CoAuthor uses a life science specialized GPT with retrieval over the documents a writer selects, plus eCTD templates and structured content authoring. Certara.AI can also run a customer's own model, so the engine depends on the deployment.

Those choices separate molecular prediction from language generation and from mechanistic simulation. Model versions, the generative model's training data and a shared update policy are not given, so a past customer analysis cannot be tied to a frozen model.

BB on Model Supply Chain DisclosureSubstantial partial disclosure, or a chain that is structurally short: an in house build, a cleared model that cannot be quietly swapped, or a deployment where the transfer does not occur at all. Naming the model provider exits the band below into this one; the axis rises from here on the completeness of the party list and on the terms that govern data once it arrives.
Vendor Published

Processing paths differ by product. Libra runs on Azure, CoAuthor gives each organization its own model, and Certara.AI can sit inside customer infrastructure. The September 2026 subprocessor list names AWS and IBM Cloud, Anthropic for enterprise writing and analysis, Cursor for software development and Microsoft for productivity applications.

Those corporate uses do not show that Claude or Copilot generates CoAuthor output. The list stops short of mapping each product to its inference provider, retention terms or a change in the model serving a customer, so the private model design gives partial visibility into who handles prompts and drafts rather than a full chain from source document to finished passage.

CC on Clinical and Operational EvidenceNamed customers, or vendor reported percentages with no method, denominator or reference standard. Scale of use is recorded here and is not treated as evidence of benefit.
Vendor Published

Libra's accuracy claim reaches 85%, without a test count, data split or full reference standard behind it. Its retrospective AZD1979 example traces to a Bayesian model built before the September 2025 commercial release, so it describes earlier work rather than today's pipeline.

CoAuthor material cites a 30% cut in first draft time, and Certara credits TFL Studio, the Phoenix Cloud module that AI PK Reports builds on, with halving the time to build tables, figures and listings. Neither figure comes with a reproducible evaluation tied to the current model. Simcyp's publication record and the 33,000 articles citing Certara products concern other methods and do not test generative writing or liver risk prediction.

BB on AI Safety and PHI StewardshipCategorical commitments are published, such as no training on customer data, without the retention schedule behind them.
Vendor Published

Each organization gets its own CoAuthor model, with retraining and active learning off by default, so ordinary use stays separate from a training workflow a customer would have to switch on. The authoring design strips identifiers automatically and reviews each passage against its source, and the Veeva RIM link lets writers reference source files in place rather than importing copies into CoAuthor.

Retention follows purpose rather than a fixed schedule for prompts, generated narratives, backups and model artifacts, and the privacy notice lets some information stripped of identifiers remain. Safety event handling and downstream deletion across Libra, Certara.AI and AI PK Reports are not described in the same detail as the authoring controls.

Regulatory and Compliance
CC on HIPAA and BAA PostureA status is stated but not supported: compliance is claimed, or business associate status stated, without the underlying document, including where the only privacy notice published covers the website rather than the service that handles patients. The mirror case grades here as well: a substantive privacy document that reaches the service, with no statement of business associate status anywhere and no scope position taken. One of the two elements the band above asks for is present, which is a grade below it and not a grade at the floor.
Vendor Published

Clinical study documents and research datasets can carry patient information, and the privacy notice covers health and medical information used in drug development, epidemiology and market access work. Processing on a customer's behalf falls under separate licenses, service agreements and data processing agreements, outside the notice's controller terms, and privacy requests about customer data go back to the customer.

No business associate statement or readable agreement sets out a US protected health information posture for CoAuthor or the wider AI suite. International transfer safeguards cover cross border movement rather than that contract scope, and CoAuthor's automatic removal of identifiers reduces identifiable content in drafts without standing in for those terms.

AA on Security Certifications and Trust CenterCertifications named with their type and version and presented as retrievable artifacts, usually through a trust portal a buyer can open without asking.
Vendor Published

BSI certificate IS 769734 covers an ISO/IEC 27001:2022 information security management system, effective March 13, 2026, revised March 19 and expiring March 12, 2029. The two page certificate names Certara.AI with CoAuthor, Layar, CODEX and its data center, along with Chemaxon, D360, Integral, Phoenix, Simcyp and other suites. Registered sites include the Pennsylvania head office and a Boston colocation data center.

The statement of applicability sits behind access control. Libra is not named in the product list, so the certificate's scope stays with the listed suites, and it covers the management system and specified operations rather than the accuracy of any model output.

CC on FDA and Regulatory StatusNo device claim is made and the product is scoped accordingly. Most administrative and operational products sit here and are not penalized for it, because this axis grades the appropriateness of the positioning rather than possession of a clearance. A stated limit on use, such as research use only or a clause that the product gives no medical advice, beside marketing that sells the product for clinical decisions also sits here, with both quoted, because the buyer has to ask which one governs.
Vendor Published

Discovery scientists and regulatory teams use the AI products to assess compounds, analyze studies and draft submissions. That work supports medicine development without making the software a cleared device or an approved product. Libra fits the move toward new approach methodologies in preclinical safety, and CoAuthor helps writers prepare material for regulatory review.

Regulators' use of other Certara tools, such as ANVISA's eCTD 4.0 submissions through GlobalSubmit in Brazil, and approvals won by customer drug programs have their own scope. No device clearance attaches to these AI modules, which serve research and professional authoring rather than diagnosis or treatment at the point of care.

BB on AI Governance and Bias DisclosureA governance framework with named process behind it, such as certification to an artificial intelligence management standard, or material written for a customer own review committee to evaluate the product with. A framework covering the vendor whole portfolio counts when its scope reaches this product; what it cannot supply is the product specific evaluation the band above asks for. Published results within the one group where the performance of this product is most at risk also sit here: evaluation on the question this axis asks, short of the comparison across groups the band above requires. A medical device clearance or certification, and the quality system behind it, does not reach this band on its own; it is graded under regulatory status. A regulator reviewed plan governing how the model may change does reach it.
Vendor Published

Source selection, passage level traceability, controlled content repositories and collaborative quality review give regulatory teams a defined way to check CoAuthor drafts. Private models with retraining off by default limit how one project's content can shape another, and the Veeva RIM link keeps writers working from the regulated source documents of record. Certara.AI can run an organization's own model, which moves part of the evaluation duty to the customer.

No current evaluation shows how the AI suite performs across therapeutic areas, document types, chemical classes or patient populations. The ISO certification covers security management rather than model behavior, so it does not fill that scientific and language evaluation gap.

CC on AI Liability and RecourseMechanisms exist that let someone challenge an output, such as audit trails, source traceability or review before commit, with nothing standing behind the output and no route for the harmed party.
Vendor Published

A writer can open the source behind any CoAuthor passage, revise it, rerun the generation and keep document versions for quality review, and scientists read Libra predictions as candidates before acting on them. Errors can be challenged inside the customer's own workflow.

The corporate privacy process takes complaints and corrections about personal information, but requests about customer research data go back to the customer. No standard accuracy warranty, indemnity for an AI error or duty to correct a wrong prediction comes with the product descriptions, and the website's legal disclaimers are general terms rather than the negotiated software agreement that would carry any such remedy.

Integration and Deployment
BB on EHR and Interoperability DepthNamed systems with read access or one directional writing, or standards support with named deployments behind it.
Vendor Published

Interoperability centers on research and regulatory systems rather than patient chart exchange. CoAuthor works inside Microsoft Word and links source files from Veeva RIM without importing them, D360 connects with Scilligence RegMol and CDD Vault, and Chemaxon tools share chemical drawing and registration across the portfolio. AI PK Reports loads material from Integral and saves completed reports back as savepoints, and the Integral MCP pilot gives external agents controlled retrieval.

These are concrete research system connections. None amounts to production read and write integration with a major EHR, and portfolio connectivity is a different capability from clinical record exchange.

BB on Deployment Model and Data ResidencyOptions and residency are stated with isolation or the processing path left open.
Vendor Published

Certara.AI is offered as a hosted cloud service or inside an organization's own infrastructure. CoAuthor uses a separate model and data store for each organization, Libra is an Azure hosted service, and Phoenix Cloud, home to TFL Studio and AI PK Reports, is hosted as well. The ISO certificate lists a Boston colocation data center for the covered Certara.AI environment.

That location is not a residency commitment for every product or customer. The privacy terms allow international processing under transfer safeguards, and the model, backup and support access path is not mapped by region. Customer managed installations narrow those boundaries, while the hosted products carry no single residency promise.

Commercial
BB on Commercial TransparencyA price or a pricing basis is published without full tiers, so a buyer can size the cost before making contact.
Vendor Published

Libra licenses run $9,000 a year for one user of the entry edition or $19,000 for one Max user, with each extra user at $2,500 a year. The entry edition carries the QSAR and Bayesian models, and Max adds batch testing, data imputation and generated reports. A free prototype evaluation is arranged through sales rather than sold as a self service plan.

Those figures cover Libra only. CoAuthor, Certara.AI, Chemaxon, Phoenix Cloud and consulting are quoted separately, as are implementation services, so one AI module carries a public price while the broader package is scoped deal by deal.

CC on Setting and Specialty CoverageCoverage is claimed broadly without specifics, or stated clearly with nothing validating it yet.
Vendor Published

Toxicology groups at pharmaceutical and biotechnology companies use Libra in early candidate selection, starting from structure and adding assay data as it arrives. Regulatory writers use CoAuthor for patient narratives, study reports, protocols and related submission content, and pharmacokinetic scientists draft reports from Phoenix output with AI PK Reports. Certara.AI reaches discovery analysis and trial data, with D360 and Chemaxon serving research informatics.

Certara counts 2,600 companies and 94,000 users across the wider portfolio, mostly for conventional biosimulation and analysis. Validation of the current AI releases across therapeutic areas and document classes is not reported. Since May 2026, the in house medical writing services team of about 220 people has belonged to Veristat, while CoAuthor stays with Certara.

Alternatives

Alternatives to Certara

Other products that do the same kind of work as Certara, and what each one does.

Peer AI provides regulatory document authoring software for pharmaceutical companies, biotechnology developers, and contract research organizations.

Cellarity is a Flagship Pioneering company in Somerville, Massachusetts, founded in 2019, that designs drugs against the behavior of a whole cell rather than against a molecular target.

Scoop provides a hosted workspace for preparing and maintaining investigational new drug submissions. Regulatory affairs teams upload scientific reports, chemistry and manufacturing packages, and supporting documents.

XtalPi is a drug and materials discovery company founded in Boston in 2015 by three MIT postdoctoral physicists, headquartered in Shenzhen and listed on the Hong Kong Stock Exchange (2228) since 2024.

Variational AI develops Enki, a generative model for designing small molecules against a research team's target product profile.

insitro is a South San Francisco AI therapeutics company founded in 2018 by Daphne Koller, built on causal biology.

Commercial

Pricing

Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.

Entry Price Pricing Basis BAA Tier Implementation Source
Libra Basic: $9,000/year for one user
$9,000 baseline
Libra's entry edition costs $9,000 a year for one user, and Libra Max costs $19,000 a year. Each additional user costs $2,500 a year. These annual licenses cover the named Libra edition, not the wider Certara portfolio. No standard business associate agreement accompanies any product edition. No implementation fee is listed. Product configuration, implementation services and consulting are quoted per engagement. Vendor Published

The entry edition includes the QSAR and Bayesian models. Max adds batch testing, data imputation and generated reports. A free prototype evaluation is arranged through sales. CoAuthor, Certara.AI, Chemaxon and other products are sold separately with no public combined price.