Is this production-ready?

No. The README calls it a functional POC and lists prerequisites: approver labeling, threshold calibration on real history, Jev route hosting on BTP or S/4, and view authorization.

What does Jev actually decide?

It returns probabilities for supplier fraud, payment diversion, overpricing, order splitting and internal fraud, plus an atypicality score and a suggested action; ABAP maps them to release, approval or block.

What SAP version does the POC need?

SAP S/4HANA on-premise with RAP and OData V4, tested on ABAP 758 with a HANA database, plus Node.js 18 or newer for the Fiori Elements app.

Can I run the POC without an OpenRouter key?

Yes. Without a key the app uses a local simulator; the real Jev model is reached through OpenRouter, and the key stays in the server-side .env and never reaches the browser.

What data leaves SAP when a purchase order is evaluated?

Only anonymized indicators such as days, ratios, counts and flags; no supplier names, supplier codes, user IDs or order numbers are sent, and an automated test enforces this.

How much does a fraud evaluation cost?

On the first real call it measured US$0.0000913 per seven-question evaluation, about US$0.09 per thousand purchase orders, since Jev charges only input tokens.

Does the POC pass SAP's clean-core check?

The ATC variant ABAP_CLOUD_READINESS reports zero findings in the core; the only two findings sit in the ZI_POC_JEV_W_CHGDOC wrapper that reads change documents, for which SAP offers no released API.

CommunityJev Integrations & Adapters

SAP S/4HANA Procurement Fraud

Proof of concept that flags suspicious purchase orders in SAP S/4HANA: ABAP computes risk indicators, a Fiori Elements app sends seven typed questions to Jev via OpenRouter, and ABAP applies thresholds to rate each order. Runs in shadow mode only.

See highlights

Open github.com

Stars: 0
Language: JavaScript
Published by: Community
Last verified: Oct 2026

seccoit/sap-s4-antifraude-jev

User repository on GitHub

View on GitHub

POC de detecção de fraude em pedidos de compra no SAP S/4HANA: indicadores em CDS/RAP, app Fiori Elements e decisão tipada do TypeSafe Jev via OpenRouter, em modo sombra.

Forks: 0
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

Prova de conceito que avalia o risco de fraude em pedidos de compra do SAP S/4HANA com o Jev, modelo de decisão tipada da TypeSafe AI, acessado pelo OpenRouter.

Quoted from the project's README. Read the full README.

Tagscommunity sap fraud-detection poc abap

Works forJev for Scoring Jev for Verification Jev for Decision gating

Used inJev for Finance & Fintech Jev for Software & DevTools

Languages: JavaScript
Platforms: Browser
Capabilities: Open sourceRequires API key

Highlights

Sends five Noul questions, one Score (1-5) and one Choice per order; ABAP, not the model, always assigns the final class
Thresholds: maximum risk below 0.20 releases, up to 0.60 routes to approval, above blocks; documented as an uncalibrated hypothesis
Measured cost on the first real call: US$0.0000913 per seven-question evaluation, about US$0.09 per thousand orders
Shadow mode only: evaluations are logged and no order is ever blocked; blocked vendors are a hard ABAP rule that bypasses Jev
Privacy: no supplier names, user IDs or order numbers leave the system, only days, ratios, counts and flags; enforced by an automated test

Watch out

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

Rot Guard

MIT-licensed Chrome extension that asks Jev whether the page, video or post you are on serves your written goals -- a category Choice, a goals Noul and a 0-4 waste-of-time Score -- then blocks or blurs in code, tightening thresholds at 2, 10 and 20 minutes of engaged time.

Jev Integrations & Adapters#community#chrome-extension#digital-wellbeing

Community

plusminushalf/rot-guardJavaScript

AutoGPT

187.5k

Autonomous agent platform with 187,504 stars for building and running agents from plain English or a visual builder. Its platform features first-class TypeSafe Jev decision blocks for typed routing, filtering, scoring and confidence-gated next-action dispatching.

Jev Integrations & Adapters#community#autonomous-agents#platform

Community

Significant-Gravitas/AutoGPTPython

x402check

x402 risk-check provider that scores agent counterparties before payment: OFAC, phishing and drainer feeds, transaction simulation, drainer-kit fingerprints and a Jev model reading context for injected instructions, issuing an ES256-signed attestation per verdict.

Jev Integrations & Adapters#community#x402#payments

Community

caiovicentino/jev-risk-check-providerTypeScript

JIT-JEV Context OS

Epistemic runtime for autonomous coding agents that uses Jev as a sub-100ms System 1 gate to shadow-evaluate capsule facts, propose evictions before bloat and prefetch memory between turns. Its fixture run reports 31.3% fewer turns and 52.6% fewer blind discovery operations, with a fail-open circuit breaker.

Jev Integrations & Adapters#community#agents#context-engineering

Community

wojciechwiesner/jit-contextPython

Vibefilter

Filament plugin that filters admin table rows by a plain-English statement: each row gets a Jev Noul probability that the statement is true, and rows at or above the threshold stay. Scores are cached by content, so reruns of the same statement are instant and free.

Jev Integrations & Adapters#community#laravel#filament

Community

vibefilter/filamentPHP

r2r-jev

Agent governance integration where Jev judges two Noul checks per tool call (beyond scope, destructive), an admission policy decides which judgments may become evidence, and admitted evidence degrades Trust, Delegation and Authorization in persistent R2R relation state.

Jev Integrations & Adapters#community#governance#agent-security

Community

Thneoly/r2r-jevRust

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai