[DE] Annex IV[DE] Technical documentation from live systems

[DE] Stop losing enterprise deals to a compliance questionnaire.

[DE] Evidence-based Annex IV documentation generated from your live ML systems — so your security review doesn't stall your sales cycle.

[DE] Run the free readiness check[DE] Join the waitlist

[DE] Reads your stack via REST & AST[DE] No model weights leave your infra[DE] EU-hosted

modeldocs scan · candidate-rankerlive
[DE] The deal blocker

[DE] Your buyer's security team asked for your Annex IV technical file. You have a spreadsheet.

[DE] What you sent

[DE] A self-attested questionnaire and a slide that says "AI Act: in progress."

// status: blocked in security review

[DE] What closes the deal

[DE] A complete Annex IV technical file where every claim points to evidence.

// status: passed, contract signed
6–11 Wochen

[DE] is what a stalled high-risk-AI security review adds to an enterprise sales cycle.

[DE] The difference

[DE] Assertion-based vs. evidence-based.

[DE] Everyone else turns a questionnaire into a PDF.

[DE] Questionnaire tools[DE] Self-attested
Q14[DE] Does the system undergo validation and testing before deployment?
YesNoPartially
Q15[DE] Describe your design specifications and architecture choices.
"[DE] We follow standard ML best practices…"
[DE] Generated output: "[DE] The provider asserts that validation and testing are performed."
modeldocs[DE] Evidence-linked
$ modeldocs map --section 2g
 
✓ §2(g) validation and testing
evidence:
eval/holdout_metrics.json
auc=0.834 recall@10=0.71
fairness/ selection_rate by group
source:
→ mlruns/9f2c1a/metrics
→ registry://candidate-ranker/3
verified: 2026-05-02 · re-scan on retrain
[DE] Each line resolves to an artefact a reviewer's engineers can open and check.
[DE] How it works

[DE] Connect once. Map automatically. Stay current.

[DE] The scanner reads metadata and source — never weights or training data.

SCHRITT 01

[DE] Connect

[DE] Point the open-source scanner at your stack.

MLflow RESTModel registryPython AST
SCHRITT 02

[DE] Map

[DE] Every artefact is matched to the Annex IV section that requires it.

§2(b) design§2(g) validation§3 monitoring
SCHRITT 03

[DE] Maintain

[DE] Models retrain and documentation decays.

[DE] Free, no signup

[DE] Annex IV Readiness Check

[DE] Nine questions about your ML stack and deployment.

[DE] Run the readiness check
9 [DE] questions~2 min [DE] to complete0 [DE] data stored
65/100
[DE] sample readiness score
[DE] Do you track experiments in MLflow or similar?
[DE] Are validation metrics versioned with the model?
[DE] Is there a documented post-market monitoring plan?
Für wen

Entwickelt für EU-KI-Anbieter gemäß Anhang III.

// primäre nutzer

Technische Gründer, CTOs & ML-Verantwortliche in EU-SaaS-Unternehmen mit 20–100 Mitarbeitenden.

Insbesondere HR-Tech- und Recruiting-Plattformen, die Anbieter von Hochrisiko-KI sind — automatisierte Vorauswahl, Rangfolge und Zuordnung von Bewerberinnen und Bewerbern. Sie sind technisch versiert, haben wenig Zeit und keine Geduld für Compliance-Theater.

Anhang III · §4(a) Beschäftigung & Personalverwaltung
Sie liefern KI, die Bewerber vorauswählt, einstuft oder zuordnet.Damit fallen Sie eindeutig in die Hochrisiko-Kategorie.
Ihre Enterprise-Kunden führen eine Lieferanten-Sicherheitsprüfung durch.Und diese umfasst jetzt auch Ihre technische Dokumentation nach dem KI-Gesetz.
Sie verbinden lieber ein System, als ein Formular auszufüllen.Nachweise aus Ihrem Stack schlagen eine 60-seitige Selbstauskunft.
[DE] EU-hosted[DE] Data stays in the EU
[DE] Open-source scanner[DE] Read and audit the code
[DE] Your code never leaves your infra[DE] Reads metadata via REST & AST only
[DE] Early access

[DE] Get your technical file out of the way.

[DE] Join the waitlist and we'll onboard EU AI Providers in cohorts.

[DE] No spam. One email when your cohort opens.

[DE] We store your work email to manage waitlist access. No data is shared with third parties.