Skip to main content

Ketryx runs design controls inside your dev tools. Seal runs design, build and quality on one record.

Seal, the GxP execution system, is your eBR, your QMS and more. Neil, your AI agent, does the paperwork across your systems. You sign. Paradromics cut operator paperwork by 80%. Start with one process beside Ketryx.

Book a demo

Where Ketryx falls short.

From Ketryx’s own pages.

  • Requirements
    In Jira or Jama
    Code and tests
    In GitHub and TestRail
    Ketryx
    “It orchestrates them.”
    Seal
    Design, build and quality on one record

    It orchestrates the tools you have.

    Ketryx says it “doesn’t replace your systems. It orchestrates them.” Requirements, code and tests stay in Jira, GitHub, TestRail and the rest. In Seal, design records sit on the same platform as each unit’s build record, quality and training.

    Ketryx, home page
  • Pricing page
    “FDA 21 CFR Part 210, 211 and 820”
    Device history record
    On no capability page
    Batch record
    On no capability page
    Seal
    Each unit’s build record

    No record of the build.

    Ketryx’s pricing page lists “FDA 21 CFR Part 210, 211 and 820”, yet no capability page describes a device history or batch record. In Seal, each unit’s build record links to the versions and lots it was built from.

    Ketryx, Pricing
  • Jira, in Ketryx’s words
    “not designed for regulated development”
    Ketryx adds
    Signatures, release gates, immutability
    Validating Jira
    By Ketryx, as a service
    Seal
    Records and checks on one platform

    Its controls rest on the tools it connects.

    Ketryx says “Jira was not designed for regulated development”. Ketryx adds Part 11 signatures and release gates inside it, makes approved items immutable, and validates Jira as a service. In Seal, records and their checks are on one platform.

    Ketryx, Jira connector

See it in Seal.

Real Seal screens with fictional test data: validating a LIMS update, from its requirements to a failed test and what is still missing.

  1. 1

    Requirements, by risk.

    Each requirement for a LIMS update carries its risk and the tests that cover it on the version under test.

    Requirements in Seal: six requirements for a LIMS, from a complete audit trail to backup and restore, each with its risk and the tests that pass on version 4.3.0.
    1. Requirements as recordsEach one published, with its own version.
    2. Risk on every requirementHigh, medium or low.
    3. Coverage, by versionWhich tests pass on 4.3.0.
  2. 2

    A failed step, with its evidence.

    The user acceptance run records what should happen, what did, and the screenshot that shows it.

    Step 3 of user acceptance run UAT-01 on version 4.3.0 in Seal: the expected out-of-specification banner, the actual result accepted without a flag, the step marked Fail, and the screenshot of the LIMS result entry attached as evidence.
    1. What should happenAn out-of-specification banner, and the result flagged.
    2. What did7.46 shown as 7.5 and accepted, with no flag.
    3. The evidence, attachedThe screenshot from the run, kept with the step.
  3. 3

    What is still missing, in one place.

    The change to the LIMS shows the required check that failed and the sign-offs still to come.

    Completion checks in Seal for change CC-003 to SampleTrack LIMS 4.3.1: two sign-offs remaining on the change and two on run UAT-01, and the required check that approved test cases pass marked Failed because the run on 4.3.1 is not finished.
    1. Sign-offs, countedTwo remaining on the change and two on the test run.
    2. A required check, failedApproved test cases must pass; the run on 4.3.1 isn’t finished.

What Seal’s customers measured.

Each figure links to the customer’s story.

QualityAcclinate and Skye Biologics

−50%
quality cycle times at Acclinate, from a procedure change to everyone trained.Acclinate
−40%
non-conformances at Skye Biologics, with training checked before execution.Skye Biologics

RecordsMiAlgae and Bioventix

−75%
documentation time at MiAlgae, from the lab to commercial production.MiAlgae
−80%
document admin time at Bioventix, after two decades on paper.Bioventix

Ketryx and Seal, side by side.

Every source is listed under References.

KetryxSeal
What it isAI-powered lifecycle management over your engineering tools¹The GxP execution system: lab, manufacturing and quality on one platform
Data modelItems extracted from connected tools, each with a Part 11 record, in a live trace matrix²Primitives your team composes: record types, their relationships, steps and checks
AIAgents for change impact, conflicts, test coverage and complaints, acting only on approval³Neil, Seal’s AI agent, prepares changes for your team to verify and approve
Runs the workEnforces SOPs and gates releases inside Jira, GitHub and other connected tools⁴Lab, manufacturing and quality work; for code, a GitHub app in beta that approves or closes pull requests
After design transferComplaints and CAPA; no device history or batch record on any capability page⁵¹Each unit’s build record links to the versions and lots it was built from
Test evidenceAutomated tests traced from Git, TestRail and Xray; an enriched SBOM and vulnerability triage⁶⁷Files from your pipeline on the change record; Seal has no SBOM or vulnerability scanning
SignaturesPart 11 signatures inside Jira; an IQ/OQ validation package provided⁸⁹Part 11 and Annex 11 signatures and audit trails; you validate your intended use
Getting startedFree for pre-market companies that have raised less than $2 million; other plans through sales¹⁰A 48-hour evaluation of one process, once the scope is agreed and the inputs received

Summary

The story
Ketryx is built for regulated teams who want compliance evidence to come out of the tools their engineers already use.¹ Erez Kaminski, who had led AI and machine learning for Amgen’s medical device division, founded it in 2021 with Jan Pöschko from Wolfram Research, and began with integrations for Jira and GitHub: “a central orchestration layer” for a single, traceable software development lifecycle.¹¹ It still works that way. Requirements, code and tests stay in Jira, Jama, GitHub and TestRail, while Ketryx keeps the trace matrix, generates the design history file and gates releases until procedures are followed.⁹⁴ Ketryx says it is trusted by 4 of the top 5 medical device manufacturers and reduces documentation burden by up to 90%.¹² Its AI agents, launched in May 2025, look for change impact, conflicting requirements and test gaps, and take no action until a person approves.¹³³ If your engineers design in those tools and someone else builds the device, it can work well. A team that also builds the device needs each unit’s build on the same records as the design.

1What Ketryx offers.

Ketryx Corporation has offices in Cambridge, Massachusetts, and Vienna, Austria.¹⁴ It was founded in 2021,¹¹ and in September 2025 it raised a $39 million Series B led by Transformation Capital, bringing its total funding to over $55 million.¹⁵ It calls its product “AI-powered lifecycle management for regulated teams”.¹

Ketryx connects to Jira and Jira Data Center, GitHub (repositories, pull requests and GitHub Actions), GitLab, Azure DevOps, Bitbucket, TestRail, Xray, Jama and Polarion, and to security tools such as Black Duck, Snyk and SonarQube.⁶ From each connected tool it extracts the items, keeps a Part 11 record and audit trail of their changes, and compiles a configurable Design and Development File.² It enforces procedures inside those tools: guardrails stop a product from progressing until it is compliant, V&V tests do not run for uncontrolled requirements or unapproved test cases, and approved items become immutable.⁴⁸ It turns scanner output into a documented SBOM, brings threat models in as traceable items, and says Medtronic cut its vulnerability triage time by 80%.⁷ Its eQMS adds document control, training, CAPA and complaint handling,⁵ and for electromechanical devices it traces hardware specifications, firmware and software to the same requirements and risks.⁹ Ketryx says Beacon Biosignals, which combines wearables with AI analysis of brain data, cut its release cycle from four weeks to two.¹⁶ There is a free plan for pre-market companies that have raised less than $2 million; other plans are priced through sales, and validation evidence is provided on request.¹⁰

For a software team, this goes well beyond Seal. Seal’s GitHub app, in beta, ties a pull request to a Seal change record and its Part 11 approval, but it does not inspect code, run tests, merge or deploy. Test evidence arrives as files from your pipeline, and Seal has no software bill of materials or vulnerability scanning.

2Running Seal alongside Ketryx.

You can start without replacing Ketryx: your software teams keep their trace matrix and release gates in it while the first process moves to Seal, such as the build record of one device. Seal has no dedicated Ketryx connector. Where a system is not on Seal’s integrations list, Seal’s REST API and Scripts can connect it.

Agree the records, the direction of transfer, the credentials and the failure handling before implementation. Start with read access, and treat any write in either direction as a separate decision with defined actions, review gates and audit evidence. Qualify the connection by defining identifiers and field mappings, testing missing, delayed and duplicate records, and reconciling transferred records against the source, as the regulatory guide describes.

Where code changes reach production, Seal’s GitHub app creates a Seal change record from a configured template when a matching pull request opens, writes a commit status that links to the record, and approves or closes the pull request when the record is approved or rejected. Ketryx can also gate deployment through GitHub,⁴ so agree which system holds the approval before adding Seal’s app to the same repository.

3Moving from Ketryx to Seal.

Move one process where the design meets the build first. In Seal, design inputs, outputs, verification, validation and design reviews are linked records, and the traceability matrix is a view of those links, so a requirement with no test shows as a gap while the design is developed. Risk controls link to the requirements and verification that implement them, and the design history file is organised from records kept during development. Each unit’s build record links to the versions and lots it was built from, so a design change shows the units it reaches, and a complaint opens on the unit’s build. The design control blueprint and the medical device blueprint set this out. Paradromics, which makes a brain-computer interface, moved from paper and legacy systems and cut operator paperwork by 80% and engineers’ traveler review time by 90%.

  1. Choose a change with consequences beyond the software, such as a firmware or component revision that reaches verification, risk controls and units already built.
  2. Bring the affected requirements, the risk file, the verification evidence and the build records. Start with sample records, or arrange an NDA before sharing confidential ones.
  3. Neil prepares the trace structure and the change workflow from your procedure, for your engineering and quality teams to verify against the device. Seal offers a 48-hour evaluation of one process, once the scope is agreed and the inputs received. Production timing then depends on scope, interfaces, migration and your release requirements.
  4. Exercise a change before relying on it: a revised requirement, the verification and risk controls it reaches, and the units built to the earlier version.
  5. Keep the design history files and release documents Ketryx generated retrievable for their retention period. Imported evidence keeps its original version and context, and importing it does not establish that it covers the current use.

References

  1. 1AI-Native Continuous Compliance for Regulated Development, Ketryx. https://www.ketryx.com/product (read 13 November 2026)
  2. 2Medical Device Total Product Life Cycle (TPLC) Documentation, Ketryx. https://www.ketryx.com/capabilities/documentation (read 13 November 2026)
  3. 3AI Agents, Ketryx. https://www.ketryx.com/capabilities/ai-agents (read 13 November 2026)
  4. 4Medical Device Quality Management System (QMS) Software, Ketryx. https://www.ketryx.com/capabilities/enforcement (read 13 November 2026)
  5. 5Medical Device Electronic Quality Management System (eQMS), Ketryx. https://www.ketryx.com/capabilities/eqms (read 13 November 2026)
  6. 6Integrations for Regulated Product Development, Ketryx. https://www.ketryx.com/capabilities/integrations (read 13 November 2026)
  7. 7Medical Device Software Bill of Materials (SBOM) Management Software, Ketryx. https://www.ketryx.com/capabilities/sbom-software-bill-of-materials (read 13 November 2026)
  8. 8FDA, EU MDR, & ISO Medical Device 62304 Compliance Software for Jira, Ketryx. https://www.ketryx.com/capabilities/jira-connector (read 13 November 2026)
  9. 9Ketryx for Medical Devices, Ketryx. https://www.ketryx.com/industries/medical-devices (read 13 November 2026)
  10. 10Pricing, Ketryx. https://www.ketryx.com/pricing (read 13 November 2026)
  11. 11About Ketryx, Ketryx. https://www.ketryx.com/about/company (read 13 November 2026)
  12. 12AI-Native Platform for Safety-Critical Products: FDA, EU MDR & ISO Compliance, Ketryx. https://www.ketryx.com/ (read 13 November 2026)
  13. 13Ketryx Launches First Validated AI Agents to Accelerate Safety-Critical Product Development, Ketryx, 1 May 2025. https://www.ketryx.com/press-release/ketryx-launches-first-validated-ai-agents-to-accelerate-safety-critical-product-development (read 13 November 2026)
  14. 14Imprint, Ketryx. https://www.ketryx.com/imprint (read 13 November 2026)
  15. 15Ketryx Raises $39M Series B Round to Power the Future of Regulated Artificial Intelligence for Life Sciences, Ketryx, 4 September 2025. https://www.ketryx.com/press-release/series-b (read 13 November 2026)
  16. 16How Beacon Biosignals Reduced Their Release Cycle to Two Weeks for AI/ML Products with PCCPs, Ketryx. https://www.ketryx.com/case-studies/beacon-biosignals (read 13 November 2026)

AQuestions and answers

Is Seal better than Ketryx?

Yes. Ketryx is built for regulated teams who want compliance evidence to come out of the tools their engineers already use. Erez Kaminski, who had led AI and machine learning for Amgen’s medical device division, founded it in 2021 with Jan Pöschko from Wolfram Research, and began with integrations for Jira and GitHub: “a central orchestration layer” for a single, traceable software development lifecycle. It still works that way. Requirements, code and tests stay in Jira, Jama, GitHub and TestRail, while Ketryx keeps the trace matrix, generates the design history file and gates releases until procedures are followed. Ketryx says it is trusted by 4 of the top 5 medical device manufacturers and reduces documentation burden by up to 90%. Its AI agents, launched in May 2025, look for change impact, conflicting requirements and test gaps, and take no action until a person approves. If your engineers design in those tools and someone else builds the device, it can work well. A team that also builds the device needs each unit’s build on the same records as the design. Seal has no fixed model of one kind of work. It is built from primitives: record types, the relationships between them, and the steps and checks that use them, which your own team composes into your process, with Neil able to prepare each change for approval. So one platform runs process development, the laboratory, manufacturing, batch release and the quality system, and each change becomes a new version, released once its selected tests pass, while earlier runs keep theirs. Moving from paper and legacy systems, Paradromics cut operator paperwork by 80% and engineers’ traveler review time by 90%.

Who makes Ketryx?

Ketryx is made by Ketryx Corporation, which has offices in Cambridge, Massachusetts, and Vienna, Austria. Erez Kaminski and Jan Pöschko founded it in 2021, and in September 2025 it raised a $39 million Series B led by Transformation Capital, bringing its total funding to over $55 million.

Is Seal an alternative to Ketryx?

Partly. Ketryx is built for software and device engineering teams that want design-control evidence to come from the tools they already use, and there it does more than Seal: more integrations, a software bill of materials and release gates inside those tools. Seal is the GxP execution system: its design records sit on the same platform as the build record of every unit, the quality system and training, and Neil prepares the changes. A team deep in Ketryx can keep it and start in Seal with one process.

Does Ketryx cover manufacturing, such as device history records or batch records?

Not on its own pages. Ketryx’s capability pages cover requirements, traceability, risk, documentation, SBOM, enforcement and an eQMS with document control, training, CAPA and complaints. None describes a device history record or a batch record, although its pricing page lists 21 CFR Parts 210, 211 and 820. In Seal, each unit’s build record links to the versions and lots it was built from, so a design change shows the units it reaches, and a complaint opens on the unit’s build.

How does Ketryx’s AI compare with Neil?

Ketryx launched its AI agents in May 2025. They look for change impact, duplicate and conflicting requirements, test coverage gaps, incomplete change requests and anomalies, triage complaints and run nightly quality reviews, and they take no action until a person approves; Ketryx’s MCP server lets Claude, ChatGPT and other AI tools query the trace matrix. Neil prepares changes to records, workflows and their checks for your team to verify and approve under change control. It finds requirements without verification and assembles the records a change reaches. It cannot approve, sign or publish, it is not used in GMP execution, and customer data trains no AI.

Does Seal work with GitHub and Jira?

With GitHub, in a narrower way than Ketryx. Seal’s Change Control GitHub App, in beta, creates a Seal change record from a configured template when a matching pull request opens, updates it as the pull request changes, writes a commit status that links to the record, and approves or closes the pull request when the record is approved or rejected. Seal holds the approval and its Part 11 signatures; merge and deployment stay in your delivery workflow. GitHub issues with a label you choose are kept as Seal records, for example as anomalies, and Seal’s GitHub Actions attach a snapshot of the code and the release’s files to the change record. The app does not inspect code, run tests, merge or deploy, and Seal has no software bill of materials or vulnerability scanning. There is no two-way Jira sync; Neil can read Jira through a connection a user grants. There is no GitLab, Bitbucket or Azure DevOps app; other tools connect through Seal’s REST API and Scripts. Ketryx connects Jira, GitHub, GitLab, Azure DevOps, Bitbucket, TestRail, Xray, Jama and Polarion, builds an enriched SBOM and gates releases inside those tools. For a software team’s lifecycle evidence, that is well beyond Seal’s GitHub app.

Can we keep Ketryx and start with Seal?

Yes. Keep Ketryx for your software teams’ trace matrix and release gates, and start in Seal with one process, such as the build record of one device. Seal has no dedicated Ketryx connector; other tools connect through Seal’s REST API and Scripts. If Ketryx already gates your pull requests, agree which system holds the approval before adding Seal’s GitHub app to the same repository.

How long does it take to start with Seal?

Seal offers a 48-hour evaluation of one process, once the scope is agreed and the inputs received. Production timing then depends on scope, interfaces, migration and your release requirements.

Start beside Ketryx.

Keep Ketryx where it serves your software teams, and start in Seal with one process, such as the build record of one device. Other tools connect through Seal’s REST API and Scripts, so nothing has to move on day one.

Book a demo