Use case
KYC and AML screening for fintech and payments
A fintech or payments firm can screen every new customer with one API call. AdverseMe checks sanctions, PEP status, adverse media and ownership across 117 sources, returns a score from 4 independent AI judges that is reproducible for 14 days, and posts a webhook when the result is ready. Every search stays on the record.
- Start a screening
- POST /v1/screen
- Auth
- X-API-Key header
- Result
- Webhook or GET
- Score
- Reproducible 14 days
The workflow, in plain language
Onboarding is the first screening. The book you already have is the bigger one.
- 1
Onboarding
Screen the applicant and, for a business account, its directors and owners before the account is opened. A clean result lets the flow continue; a hit routes to a reviewer.
- 2
Periodic review
Customers already on the books need re-screening on a schedule, and immediately when a sanctions list changes.
- 3
Merchant and counterparty checks
Payment processors screen merchants and their beneficial owners, where a hidden owner is the risk rather than the named signatory.
- 4
The audit question
When a regulator or partner bank asks what you checked and why you decided as you did, the answer has to be on file, per customer.
How AdverseMe fits
Screening as an API call, with the working kept
One call, every source
POST /v1/screen with an X-API-Key starts a person or company screening across 117 sources, including 13 official sanctions feeds snapshotted twice a day. Poll GET /v1/screening/{id} or receive a webhook.
API reference →A score you can put in a rule
4 judges from 4 model families score blind at temperature zero. The median is the score, dissent is counted, and the same evidence returns the same score within 14 days, so a threshold in your onboarding logic behaves the same way twice.
AI council →Ownership without a second vendor
Company screenings trace officers, parents and beneficial owners to depth 6 with GLEIF LEI and SEC 13D and 13G edges. The OFAC 50 percent rule is computed for you.
Ownership graph →Delta re-screening for the whole book
Put customers on a daily, weekly, monthly monitor. Light delta checks against changed list versions are free; only a confirmed hit triggers a full re-screen and uses 1 credit.
Monitoring →Evidence per customer
GET /v1/screening/{id}/pdf returns the branded report with every judge's reasoning and the Engine Trace. Attach it to the customer record and it is there when someone asks.
Engine Trace →Migrate the back book
Upload the existing customer list as a CSV, watch it screen with live progress, and escalate hits to cases for review.
Batch screening →What a call looks like
POST /v1/screen
X-API-Key: your_key
{
"type": "person",
"name": "Jonathan A. Doe",
"country": "GB"
}
201 Created
{ "id": "scr_...", "status": "running" }
GET /v1/screening/scr_...
GET /v1/screening/scr_.../pdf
GET /v1/creditsField names are illustrative; the API page has the exact request and response shapes. Plans from $49 a month on the pricing page.
Frequently asked questions
How do I integrate AdverseMe into an onboarding flow?
Create an API key in the app, then POST /v1/screen with the X-API-Key header and the person or company details. Register a webhook endpoint to receive the result, or poll GET /v1/screening/{id}. GET /v1/credits tells you how many screenings remain.
Is the score stable enough to automate decisions on?
Yes within its window. The 4 judges run at temperature zero with fixed seeds and the median is the score, so the same evidence returns the same score for 14 days. You can automate a clear-below threshold and route the rest to review.
Does the API cost more than the app?
No. An API screening uses the same 1 credit as a screening started in the app. Fetching results, PDFs or your balance over the API does not use credits.
Where is customer data processed?
On Microsoft Azure in Western Europe. Processing is GDPR-aligned and security practices follow SOC 2 principles. Every action is written to the audit log.
Put screening in the onboarding flow this week.
Plans start at $49 a month for 50 screenings. Every plan includes the AI council, the ownership graph, Engine Trace and PDF reports.