Castor Catalyst · For sponsors and CROs

Retyping source data is slow and error-prone. Catalyst turns it into study data, under the sponsor's or CRO's review policy.

Version 2.1 · Updated 16 October 2026

Castor Catalyst is an AI extraction layer for clinical studies. It turns worksheets, scans and electronic medical record (EMR) pages into values in the electronic data capture (EDC) system, checked under the review policy the sponsor or CRO sets.

Before redactionFictional patientLab report before redaction, showing the patient's name, IDs, address, phone, physician and dates, and the first metabolic panel results.
After redactionThe same lab report after redaction: the identifiers are blacked out, dates are reduced to month and year, and age, sex and all results are kept.
Castor Catalyst redaction engine on a lab report: identifiers removed, clinical values kept. In this example, dates are reduced to month and year.

  • A workflow, not one model

    Five stages with a check at each: intake, redaction, extraction, review and commit to the EDC.

  • Redacted before extraction

    Redaction uses no large language models, and the site approves every redaction. The AI pipeline only ever sees redacted copies.

  • You set the review policy

    100% human review today, with every action in the audit trail. Risk-based review is coming soon. Coming soon

The redaction engine finds identifiers with rules and small purpose-built recognisers trained only on synthetic documents; no large language models and no external services.

Measured results

Faster than abstraction by hand, with fewer errors

  • Minutes per chart

    • Manual abstraction39 min
    • With Catalyst6 min

    Castor internal benchmarking.

  • Data errors

    • Manual abstraction6.6%
    • With Catalyst, after reviewunder 0.7%

    Two sources: manual from a published meta-analysis of 93 studies;1 Catalyst from Castor's internal benchmark.

  • Identifiers redacted

    • Single clinician, published study81%
    • Castor Catalyst redaction engine99.5%

    Engine: Castor's benchmark. Clinician: a published study on a different document set.2 People check the redactions, and the site approves them.

  • Castor internal analysis80%

    Lower cost of chart abstraction. Castor internal analysis of 12 comparative studies.

  • Production telemetry0.8%

    Of AI-extracted values changed by reviewers, in one live paper-based study in 2026.

Catalyst in use today

Sites in the US and Europe already use Catalyst in live studies. Three anonymised studies are in the case studies.

Processed in live studies, 2026
  • 40+sites
  • 4countries
  • 350+source documents
  • 15,000+data points
Active countries
  • United States
  • Germany
  • United Kingdom
  • France

What has been measured, in full

How it works

Identifiers are redacted before any AI reads the source

Two Castor services, kept apart. The Catalyst Source Gateway redacts identifiers for the site, with no large language models. Hosted on Microsoft Azure; no other subprocessors. The Catalyst pipeline extracts values for the sponsor, from redacted copies only. The providers used in a study, such as Google for AI extraction, are in Castor's sub-processor list.

Data flow for US and EU studies; UK-hosted studies: coming soon. Patient and site: the patient consents to the study and the original record stays at the site. The site redacts a document before upload with a site redaction tool (path A) or in the browser (path B, coming soon); both arrive in the Catalyst Source Gateway as a redacted copy, shown as one arrow. Or the site uploads it unredacted and the Castor redaction engine redacts it in the Gateway (path C). The Catalyst Source Gateway is a separate service with no large language models, study-isolated, encrypted, in the site's country. Hosted on Microsoft Azure; no other subprocessors. The site approves the redactions before any upload goes to the pipeline. US only: the patient can authorise retrieval of their own health records; these participant-consented records go through the same redaction engine and skip site approval. In the Catalyst pipeline, the sponsor's service, AI extracts values from redacted copies only, then review by Castor under the review policy the sponsor or CRO sets. AI processing runs in the US for US studies and in the EU for EU studies; UK-hosted studies are coming soon, ask your Castor contact. The EDC holds no personal identifiers.
Data flow for US and EU studies. UK-hosted studies: Coming soon

  • Regions

    AI processing runs in the US for US studies and in the EU for EU studies.

    UK-hosted studies: Coming soon Ask your Castor contact.

  • No training on your data

    Castor doesn't train AI models using your data from Catalyst.

Which service does what, and where the data is

Review

Today every extracted value goes to a reviewer

Castor reviewers work under the policy the sponsor or CRO sets; the investigator remains responsible for the study data.

Who decides what enters the EDC

  • 100% human review

    A person reviews every extracted value before it reaches the EDC.

  • Risk-based review Coming soon

    Values above the confidence threshold the sponsor or CRO sets go to the EDC without a person, and the fields the sponsor or CRO picks, such as adverse events and primary endpoints, always go to a reviewer.

Compliance

What quality and IT teams will check

For the study, the sponsor is the data controller and Castor its processor.

  • 21 CFR Part 11

    US electronic records rules, and EU Annex 11. The sponsor or CRO validates.

  • ISO/IEC 27001

    Since 2017. Also ISO 9001 (since 2018) and NEN 7510 (since 2017), the Dutch healthcare security standard.

  • GxP validation

    Validation package from Castor quality assurance, on request.

What a quality or privacy review will ask

Next step

Test Catalyst on the study's own documents

Share a protocol and sample documents; synthetic samples are welcome. Castor benchmarks the extraction on them.