Skip to content
citaraby Fractal Q
ENFR
Request early access

Early access for founding partners

Every finding, traced to its source.

Citara reads your technical documentation the way a notified body will. It shows every gap and every contradiction before the audit does, each one linked to the exact page that proves it.

Request early accessSee an example finding
Technical file reviewIllustrative

Findings

3 major5 minor3 to review

MajorClaim not supported by the evidence

The instructions for use declare compatibility with CT scanners from all major manufacturers; the clinical performance study used images from a single manufacturer.

  1. [1]

    Instructions for use, section Compatibility · p. 6

    “… non-contrast head CT acquired on scanners from all major manufacturers.”

  2. [2]

    Clinical performance report, section Datasets · p. 14

    All 1,214 studies were acquired on scanners from one manufacturer, at two sites.

Checked against
MDR Article 7(a) · Annex I, section 23.4(e) · MDCG 2020-1, clinical performance
Suggested action
Either validate clinical performance on the other scanner manufacturers you claim, or restrict the compatibility statement in the instructions for use to the manufacturer and models validated.

Reads your file against

  • MDR (EU) 2017/745
  • Harmonised standards, Decision (EU) 2021/1182
  • MDCG 2019-11
  • MDCG 2020-1
  • Team-NB technical documentation guidance

A notified body reads across documents. So does Citara.

The report

Every finding ranked, sourced and ready to assign.

One view of the whole file: what blocks a review, what weakens it, and what needs an expert’s judgement. Each line opens on the pages that prove it.

Review reportIllustrative

All 11Major 3Minor 5To review 3

SeverityFindingSourcesRequirement
MajorThe instructions for use claim 95% sensitivity; the clinical performance report gives 92.1% on the external test set. 95% is the internal validation figure.IFU · p. 5Clinical performance report · p. 21Art. 7 · Annex I §23.4(e)
MajorVersion 2.4 retrains the model on new data; the clinical performance study was run on version 2.3, and no analysis shows its results still hold.Release notes · p. 1Clinical performance report · p. 3Art. 61(11) · Annex I §17.2
MajorThe risk file controls missed haemorrhage on motion-degraded scans with an image-quality warning; the summative usability evaluation never tests whether readers notice it.Risk file · p. 18Usability report · p. 11Annex I §4 · §5
To reviewOne of the two test sites also supplied training data; the report does not argue that the test set is independent.Clinical performance report · p. 9Data management plan · p. 4Annex I §17.2 · MDCG 2020-1
MinorThe instructions for use accept Windows Server 2016; the cybersecurity risk assessment covers only 2019 and 2022.IFU · p. 3Cybersecurity assessment · p. 6Annex I §17.2 · §17.4
MinorThe PMCF plan tracks sensitivity but not the false-positive rate, which the risk file names as the cause of alert fatigue.PMCF plan · p. 3Risk file · p. 22Annex XIV Part B §6.1
To reviewClass IIa is argued under Rule 11 for a triage tool; a missed intracranial haemorrhage can cause serious deterioration, which points to class IIb.Classification rationale · p. 2IFU · p. 4Annex VIII Rule 11 · MDCG 2019-11

Why it matters

A technical file is hundreds of pages, written by several teams, that must agree and must be enough.

Notified bodies read across documents and ask for the evidence behind every claim. A performance claim the clinical data do not cover, a requirement left without proof, a test report on the wrong software version: any one of them can stop a review. Finding them by hand takes weeks of expert time, every time a document changes. Citara writes each finding so both sides can act on it: the requirement for the regulatory team, the exact page and the expected correction for the engineers who own the document.

  • 31 Dec 2027End of the MDR transition for class III and implantable class IIb devices
  • 31 Dec 2028End of the transition for other class IIb, class IIa and class I devices
  • 38,000+Medical technology companies in Europe, around nine in ten of them SMEs

Sources: Regulation (EU) 2023/607; MedTech Europe, Facts & Figures 2025.

What Citara checks

Five reviews, run across the whole file at once.

  • 01

    Completeness

    Every element Annex II and Annex III expect: present, current and locatable.

  • 02

    Safety and performance requirements

    Each general safety and performance requirement linked to the evidence that supports it, or flagged when the link is missing.

  • 03

    Consistency across documents

    Intended purpose, claims, indications, patient populations and risks compared across the instructions for use, labelling, risk file and clinical evaluation.

  • 04

    Clinical evaluation

    Your clinical evaluation report checked against Article 61 and Annex XIV: the clinical data, the benefit claimed, the post-market follow-up.

  • 05

    Reference integrity

    Every cited standard and guidance document checked: the edition you name is current, harmonised when you say it is, and dated.

How it works

From upload to a report you can act on.

  1. 01

    Upload

    Your technical documentation as PDF files.

  2. 02

    Review

    Citara reads across every document at once, the way a reviewer does.

  3. 03

    Report

    Each finding has a severity, the requirement it relates to and the pages that prove it.

  4. 04

    Decide

    Your team accepts, corrects or contests each finding. Nothing is closed without you.

What Citara does not do

A reviewer at your side, not in your place.

  • It does not certify conformity. Your notified body does.
  • It does not replace your regulatory team or your person responsible for regulatory compliance. It gives them back the weeks spent cross-checking.
  • It does not decide judgement calls such as classification, equivalence or regulatory strategy. It flags them for expert review.

Built for software medical devices

Software teams ship fast. Regulation does not.

Citara starts where software manufacturers struggle most: the clinical evaluation of medical device software, the intended purpose and its consistency across documents, and, next, for software that includes AI, the layer the AI Act adds on top of the MDR.

  • MDCG 2020-1
  • Intended purpose
  • Clinical evidence
  • AI Act overlay

Confidentiality

Your technical file stays yours.

A technical file holds your know-how. These are our commitments.

  • Hosted in the European Union
  • Never used to train any model
  • Each client isolated from every other
  • Encrypted in transit and at rest
  • Data processing agreement provided

Modules

One review engine, several regulatory questions.

  • Early access

    Citara MDR

    Gap analysis and consistency review of MDR technical documentation.

  • In development

    Citara AI Act

    The AI Act documentation layer for medical software that includes AI.

  • In development

    Citara IVDR

    The same review for in vitro diagnostic devices.

  • In development

    Citara Studies

    A platform to run the reader studies that validate AI imaging software: cases shown to radiologists, readings collected, statistics and a report ready for the clinical evaluation.

How Citara is built

Built the way a reviewer reads.

Citara is developed by FractalQ Care. Every rule cites the regulatory text it comes from, and every finding the page it rests on, because that is the first thing a reviewer asks for.

Founding partners

Try Citara on one technical file.

We are opening a small number of founding-partner places: early access to the software for your own team, founding-partner terms, and a direct line to the people who build the product.

Request early access