TL;DR
Thorsten Meyer AI announced QAtrial, an open-source quality and compliance platform for regulated life sciences teams. The project is built around provenance for AI-assisted outputs, but the source stresses that users remain responsible for validation and regulatory compliance.
Thorsten Meyer AI has announced QAtrial, an open-source quality and compliance platform for regulated life sciences work that records provenance for AI-assisted outputs, including the model, version, purpose, human review, electronic signature and audit trail.
According to the announcement, QAtrial is aimed at GxP environments, including good manufacturing, laboratory and clinical practice settings, where validated systems, signed records and traceability are required. The platform is described as covering compliance primitives such as CAPA workflows, electronic signatures and traceability matrices.
The project is AGPL-3.0 and self-hostable, with the announcement describing support for on-premises and air-gapped environments. Thorsten Meyer AI says QAtrial is provider-agnostic, with OpenAI-compatible and Anthropic routing, and records provenance for each AI-assisted output.
The source states that QAtrial is designed to align with 21 CFR Part 11 and EU Annex 11. It also states plainly that alignment is not validation, certification or a guarantee of compliance, and that computer-system validation and regulatory duties remain with users.
AI Must Leave Evidence
The announcement addresses a central barrier to AI use in regulated life sciences: auditability. In regulated QA, teams must be able to show who created or approved a record, when it happened, what changed and why. A system that produces AI text without a traceable record can create risk rather than reduce it.
QAtrial’s stated approach is to treat the model as an attributable contributor rather than an unnamed assistant. If implemented as described, that could help teams use AI for drafting and cross-referencing while preserving human review and signed accountability.
The project also matters because it is open source and self-hostable. For regulated teams, vendor lock-in and changing model behavior can affect validation work. The announcement frames provider choice and local deployment as ways to keep regulated data and qualification processes under user control.
Software Development for GxP Regulated Industries: Deliver GxP Compliance Software in an Agile Way
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Built for GxP Records
QAtrial was introduced in Thorsten Meyer AI’s Built in Public series as Day 12 of 19 and as part of the portfolio’s Open / Reg layer. The source says the product joins Glasspane in that family and is intended to make regulated systems inspectable by design.
The announcement uses CAPA-2026-0142 as an example workflow, showing an AI-assisted draft tied to a deviation record, followed by review, electronic signature and audit logging. It also shows a traceability chain connecting a requirement, risk, test and result.
The product is positioned for work that is often slow and document-heavy, including drafting, linking requirements to tests and assembling traceability records. The source presents AI as useful for that labor, while saying rigor, review and signatures remain human responsibilities.
“You can’t put an unaccountable black box into a regulated process.”
— Thorsten Meyer AI announcement
Digital Signature Pad with LCD Screen, Electronic Signature Capture Pad for Word, Excel, PowerPoint, PDF Documents, USB Powered Signature Pad for Office, Business, Bank, Hospital, Notary
Seamless Document Signing: Supports direct electronic signature on Word, Excel, PowerPoint, and PDF files without printing or scanning,…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Validation Still Falls to Users
It is not yet clear from the source how mature QAtrial is as a deployable system, which organizations are using it, or what independent reviews have been completed. The announcement does not provide adoption figures, validation packages, audit results or production case studies.
The compliance position also remains bounded. The source says QAtrial is designed to align with 21 CFR Part 11 and EU Annex 11, but it is not validated, certified or legal or regulatory advice. AI-assisted outputs may contain errors and require qualified human review.
audit trail software for clinical research
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Qualification Evidence Comes Next
The next test for QAtrial will be whether regulated teams can install, inspect and qualify the platform in their own environments. Readers should watch for repository details, implementation documentation, validation support materials, security review information and examples of real GxP workflows.
Thorsten Meyer AI’s Built in Public series is continuing beyond Day 12, so further portfolio announcements may clarify how QAtrial connects with other Open / Reg tools.
self-hosted quality management system
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is QAtrial?
QAtrial is described as an open-source quality and compliance platform for regulated life sciences work, with features for CAPA, electronic signatures, traceability and AI-output provenance.
Is QAtrial certified for 21 CFR Part 11 or EU Annex 11?
No certification is confirmed in the source. The announcement says QAtrial is designed to align with those frameworks, but it is not validated, certified or a guarantee of compliance.
What makes QAtrial different from general AI tools?
The stated difference is provenance. QAtrial records the model, version and purpose behind AI-assisted outputs, then ties them to human review, electronic signature and an audit trail.
Can regulated companies use QAtrial as-is?
The source says QAtrial is provided as open source under AGPL-3.0 and is self-hostable, but users remain responsible for system validation, qualification and regulatory obligations.
Source: Thorsten Meyer AI
