Workspace · one team, one source space, one process
Your team works with its own sources and controlled drafts — with full content authority: you check every draft via the citation, without tearing out existing systems.
EU-hosted · no model training · you keep content authority
A controlled GxP workspace, not a chatbotsource-bound · QA-decided
Drafts are generated only from your approved sources – a closed, controlled space.
Security & standards · built for traceable evidence
Founded by CSV experts with 15+ years of domain expertise
What the Workspace delivers
The Workspace is for teams who don't want to understand traqx in theory first. Load the relevant sources, check a real question, have a draft created and see whether the answer is review-ready enough for your day-to-day work.
01 · Citation & source binding
Sources before phrasing
Answers emerge from approved sources, not from gut feeling. Every regulatory statement ends on a clickable source from your pool.
02 · Human-in-the-Loop
Authority, not autopilot
You keep content authority over every draft. The suggestion is built from sources and sits visibly alongside — you check it on its merits via the citation; nothing unverified is presented as truth.
03 · Deterministic check
Binary source check
A source reference holds or fails — no invented score. A deterministic check marks what does not hold as unverified, instead of presenting it as truth.
04 · Generate · Verify · Monitor
Your own rules first
Tenant SOPs, templates and policies take precedence over generic knowledge. The AI drafts from your rule set — conflicts between sources are made visible.
05 · Ghost Values
No model training
Customer data is not used to train models. EU-hosted, with your industry and company context — and a traceable audit trail.
The Agent Library
Pre-configured working profiles from day one — each with a fixed answer structure and source binding in every section. You keep content authority and check every section via the citation. Click through them.
Answers GxP compliance questions with precedence for tenant SOPs and approved sources. QA teams reach a robust assessment faster and see immediately where an answer is solid, limited or not sufficiently evidenced.
Supports preparation for FDA, EMA or notified-body inspections. Audit preparation becomes more focused: teams waste less time on generic preparation and see earlier where evidence is missing.
Grenze The agent does not declare audit readiness and does not guarantee an inspection outcome.
Supports risk-based validation work to GAMP / CSA logic. Validation leads gain a consistent baseline framework and reduce over- or under-validation through visible rationale.
Grenze The agent does not produce executable tests without human review and no final validation decision.
Drafts, revises and harmonises SOPs along existing templates, terminology and conventions. SOP reviews start from a cleaner state: teams see faster where content, role logic or references do not hold up.
Grenze The agent does not invent SOP numbers, form IDs or references. Missing references are flagged.
Reviews GxP documents before the QA review for compliance, quality and consistency risks. Teams catch gaps earlier, before they cost time and trust in the formal review.
Grenze The agent can prepare a sign-off recommendation, but cannot grant approval.
Every agent answers in a fixed section structure, with inline citations in every section — the disclaimer is mandatory as the last line, never skippable.
Your own agents
When your team regularly handles the same GxP questions, document types or review patterns, additional agents can be configured along your sources and rules.
Supplier evaluation
Check documents against defined qualification criteria, flag gaps, prepare follow-up questions.
Pseudo-code review
Check technical specifications, configuration logic or pseudo-code against validation requirements.
Email review
Check external communication for GxP relevance, commitments, unclear responsibilities and missing source references.
What you can do today
Today, on your own material. The Workspace carries that question source-bound through your assigned documents and returns a checked first draft — you check it via the citation and move it into your own system. Five fields, well beyond the meeting minutes.
Cluster A
Change & impact analysis
“What changes about a URS or SOP when the process or a change changes?” The right GxP agent works through the affected document set and lays out a source-bound candidate list of the affected passages — every requirement, every section with an inline citation to the triggering point. Cross-document impact instead of line-by-line handwork after every workshop.
Cluster B
Audit preparation
Which assigned documents are affected by an audit on topic X? The Audit agent delivers the source-backed scope candidate list — and derives a hypothesis-based mock-audit question catalogue from your uploaded PIC/S sources and FDA Warning Letters, each question cited to the assigned clause.
Cluster C
Draft & specification
The Workspace drafts URS, Functional Spec and SOP drafts along your templates and policies — from the assigned rule sets it knows how a URS has to look at your house. Every paragraph cited, unverified flags what does not hold: writing becomes checking. Only the finished result is moved into your own system by hand.
Cluster D
Vendor & supplier comparison
Vendor proposals against your SOPs, policies and URS: the agent lays out matches, gaps and contradictions source-backed side by side and, on conflict, presents both passages verbatim. The “fits best” call is made by the human.
Cluster E
Team standard & your own agents
Share proven workbooks and custom agents: same sources, same agent, same method. One senior's approach becomes an evidenceable team standard, scalable across colleagues and sites.
Every use case delivers a first draft or a candidate list that you check on its merits via the citation — source-dependent, without guaranteed completeness and without any autonomous decision.
Honest about maturity
We label what is not yet finished, instead of presenting it as available today.
Live today
✓ Your own rule sets
✓ Your own industry guides
✓ RAG citation
✓ Binary deterministic source check
✓ GDPR / EU hosting
✓ No model training
✓ Content authority & source check
In preparation
in preparation Excel outputs
in preparation Native Word creation
in preparation Self-service licensing
“In preparation” means: not yet in production. Until then we set up access together with you.
In comparison
What the source-checked difference is, row by row.
| traqx Workspace | Google NotebookLM | ChatGPT / Claude.ai | Microsoft Copilot | eQMS (Veeva / MasterControl) | |
|---|---|---|---|---|---|
| Source binding / Citation | ✓ Inline, from your pool | ✓ Citations | ✗ none enforced | Partly — M365, no GxP scope | ✗ Record entry, no co-author |
| Deterministic source-reference check | ✓ Anchor check (non-AI) | ✗ | ✗ | ✗ | ✗ |
| GxP role agents | ✓ 5 + custom | ✗ | ✗ | ✗ | ✗ no co-author |
| GxP domain knowledge (SOP hierarchy, conflict) | ✓ | ✗ | ✗ | ✗ | Partly — workflow, no AI layer |
| Multi-model | ✓ Opus/Sonnet/GPT/Gemini/Mistral | ✗ Gemini only | ✗ single-vendor | Partly — primarily OpenAI | ✗ |
| Audit-Trail per AI call | ✓ Prompt, retrieval, model, provider | ✗ | ✗ | Partly — IT logs | ✓ Record level |
| Formal approval workflow | ✗ not in Workspace | ✗ | ✗ | ✗ | ✓ Part-11 e-sign |
| EU region / no training | ✓ EU region · no model training | ✗ US, tier-dependent | Partly — Enterprise tier | Partly — M365-dependent | ✓ established residency options |
No single row is the moat — the combination on one Agent surface is: source-checked GxP generation right where the draft emerges. Horizontal tools have the surface without the control; the eQMS has the control, but no AI co-author.
EU region and no model training aren't an add-on — they're the basis of every Workspace licence; nothing ships without it.
Common question
Do I have to validate traqx myself?
Documentation to support your own validation, a deterministic source check, a traceable audit trail and EU hosting — without model training on your data.
The final validation and release in your own QMS remains your responsibility. traqx provides the basis but makes no regulatory decision for you.
Terms
traqx Workspace · pro Lizenz
from approx. €200 / licence / month
The Workspace starts at approx. €200 per licence and month. Validation and implementation packages are agreed separately depending on scope.
Designed for audit traceability: source binding, deterministic source check and audit trail support your own validation. Validation and implementation packages are agreed separately depending on scope.
Start with Workspace →The Workspace takes the fear out of change: no full migration, no months-long programme, no flying blind. A limited process shows whether traqx fits the way your team works — before you decide on Suite, rollout or integration.
No model training on your data. Deletion of the source space can be governed contractually. Content authority stays with your team.