FREESanctions, PEP & AML/CFT screening database. Search any name.
Autogon AI
download →
← All industries
Exchanges · Stablecoin rails · On/off-rampsApplication Shield · Omniguard AML4 min read

Security for Crypto & Stablecoin Businesses

Survive a breach and a regulator: app, API, on-ramp and AML security that complements your custody.

How Nemesis protects Security for Crypto & Stablecoin Businesses in one pass
Live flow: attacks stopped at each layer, your real users served.

There are two ways a crypto business dies: a breach, and a regulator. Both are more likely than ever, and both live on the surface Nemesis protects.

The stakes

A crypto business rarely dies of low volume. It dies in a single day. Billions are stolen from the industry every year, and exchanges that suffered a major hack have shut down. On the other side, no AML program means no license, no bank account, and no on/off-ramp, and regulators now demand the Travel Rule, sanctions screening, suspicious-transaction reporting and monthly monitoring metrics.

Our lane, honestly

Any vendor claiming to secure everything in crypto is not being straight with you. Private keys belong in an HSM or MPC, and smart contracts belong with auditors. Nemesis is precise about its lane, and it is the lane where most of your loss and all of your regulatory risk actually lives:

  • Trading app and web platform: zero-days, front-end and DNS abuse.
  • APIs and the on/off-ramp: API-key abuse, automation, scraping.
  • User accounts and auth: SIM-swap account takeover, credential stuffing.
  • Fiat rails: card fraud, fake deposits, mule accounts.
  • AML and transaction monitoring: sanctions, Travel Rule, STR filing.

It works alongside your custody and key management, not instead of them.

What Nemesis does

One agent, in your own environment: Edge stops volumetric attacks, Application Shield holds the account and API layer against takeover and abuse, and Omniguard flags sanctioned, mule and scam flows moving through your ramp and files the report. The same agent that blocks the API-key abuse also scores the withdrawal behind it.

Don't be the next headline

Scam and mule money cashes out through exchanges, and a single incident can become your regulatory and reputational death. Omniguard flags those flows before they clear, so the scandal never has your name on it.

Getting started

Start with a 30-day pilot on your platform, observe-only, in your own environment. You get a report you can hand your bank and your regulator.

Integrate Nemesis

Plug Nemesis into your stack

One agent runs all three. Start with whichever you need. Most begin with Omniguard for the money, add Application Security for the apps behind it, and Edge to keep it all standing.

See it working - live demo

A real, working app for this sector, deployed twice: one build protected by Nemesis end to end, one left unprotected. Run the same attacks on both - fraud, injection, account takeover, BOLA and mule cash-outs - and watch the protected build block what the other lets through. It opens a guided walkthrough so you can run the end-to-end tests yourself.

Open the protected demo →Open the unprotected demo →

On-ramp, off-ramp and withdrawals move value fast. Score every movement and screen every counterparty, including wallet addresses, before value leaves your platform, and meet your VASP monitoring and reporting duties.

Free to start

Screen every counterparty and the name behind a wallet against sanctions and PEP lists, free. 100 screens a day, no card, no function to build. Search it now or wire the API, then grow into full transaction scoring when you are ready.

POST https://shield.nemesislabs.xyz/api/v1/omniguard/verify   { "check":"sanctions_pep", "subject":"…" }   # free · no function needed
1 Who are you?

You build the product and own the payment code end to end.

Your access to transactions

Full code control - insert Omniguard inline on the payment path.

Recommended: Inline API / SDK
What you are fighting
Card testing & BIN attacksMule & payout fraudChargeback / first-party fraudAccount takeoverStructuring & scam payoutsSanctions & PEP screening
CBN / NFIU obligations Omniguard covers for you
✓ Real-time monitoring✓ Sanctions & PEP screening✓ STR to the NFIU (goAML)✓ KYC/CDD (BVN/NIN)✓ PCI-DSS for card data✓ PSP / switch licensing conditions
2 How do you reach your transactions?

Inline API / SDK

You call Omniguard from your payment, transfer or checkout path, before you post the entry. The verdict comes back in one request; you block, hold or allow.

LatencyReal time · sub-second
EnforcementDecline before the transaction commits
Best forFintechs, PSPs, wallets, e-commerce and digital-first banks - anyone who controls their own transaction code.
1

Create a function, get your key

A function is region × app × event (e.g. NG × payments × transfer). Create it in the Omniguard console, or via the API with your developer key. You get a function id and an ingest token (your API key). Sector rules are seeded automatically.

curl https://shield.nemesislabs.xyz/api/v1/omniguard/functions \
  -H "authorization: Bearer $DEV_KEY" \
  -H "content-type: application/json" \
  -d '{"name":"NG payments","industry":"banking","event":"transfer","region":"NG"}'
# → { "functionId":"fn_…", "ingestToken":"og_…", "rulesSeeded":16 }
2

Score every transaction before it commits

Send the transaction to /score on the payment path. Include what you have; Omniguard enriches the rest (BIN, first-time-payee, velocity, graph, breach). Any custom field you add is scored too.

curl https://shield.nemesislabs.xyz/api/v1/omniguard/score \
  -H "authorization: Bearer $OMNIGUARD_KEY" -H "content-type: application/json" \
  -d '{"function_id":"fn_…","customer_ref":"cust_913","amount":1500000,
       "currency":"NGN","channel":"web","country":"NG",
       "counterparty_account":"0123456789","beneficiary_name":"A. Bello"}'
# → { "verdict":"block", "overall_score":97, "rule_score":97, "ai_score":null,
#     "reasons":[{"signal":"Beneficiary is a flagged mule","contribution":55}],
#     "transaction_id":"txn_…", "ctr":{"reportable":false} }
3

Act on the verdict

block → decline. review → hold or step-up (OTP, call-back). allow → proceed. Test everything with dry_run:true first (free, persists nothing).

if (r.verdict === "block")  decline(txn);
else if (r.verdict === "review") holdForReview(txn);
else                            proceed(txn);
4

Screen the parties

One call, no function needed. Screen a name against sanctions/PEP, or verify identity (BVN/NIN/passport). Screening is free; KYC is metered.

curl https://shield.nemesislabs.xyz/api/v1/omniguard/verify -H "authorization: Bearer $OMNIGUARD_KEY" \
  -d '{"check":"sanctions_pep","subject":"Ibrahim Musa"}'
# → { "risk":"hit", "data":{"lists":["OFAC (US)"]}, "usage":{…} }
5

Enforce, learn, report

Flip the function to enforce so block is binding. Send true outcomes back to sharpen the model, and file goAML STR/CTR from the case.

curl https://shield.nemesislabs.xyz/api/v1/omniguard/outcome -d '{"transaction_id":"txn_…","outcome":"fraud"}'
# STR for the NFIU:  GET https://shield.nemesislabs.xyz/api/v1/omniguard/reports/{id}/goaml  (goAML XML)

Mapped to CBN & NFIU compliance

Transaction monitoring
CBN AML/CFT/CPF Regulations - ongoing, risk-based monitoring of every transaction.
Real-time /score with the fraud + AML rule engine (structuring, mule, velocity, cross-signal, graph).
Sanctions screening
Screen parties against the Nigeria Sanctions List (TPCA 2022) plus UN, OFAC, EU and UK.
/verify sanctions_pep + the Nemesis watchlist (OFAC / UN / EU / UK / PEP / EFCC).
PEP identification & EDD
Identify politically exposed persons and apply enhanced due diligence.
PEP screening flags the party and routes it to review / EDD instead of a silent pass.
Suspicious Transaction Reports
File STR/SAR to the NFIU on the goAML platform, with a valid predicate, category and indicators.
goAML-format STR export, auto-tagged and validated to survive NFIU auto-reject.
Currency Transaction Reports
Report cash at or above ₦5m (individual) / ₦10m (corporate).
Aggregated-cash CTR rule flags reportable cash; export the CTR XML.
KYC / CDD
Verify customer identity at onboarding (BVN, NIN) and keep it current.
/verify bvn / nin / passport against the identity network.
Fraud reporting & records
Report fraud (NIBSS) and keep records for five years.
Every scored transaction is logged with its reasons, source provenance and verdict - a ready audit trail.
Risk-based approach
Apply controls proportional to risk.
Calibrated 0-100 risk score with tiered review / block thresholds per function.

The whole flow in four calls

# 1 · create a function → get your key (or do it in the console)
POST https://shield.nemesislabs.xyz/api/v1/omniguard/functions   { "name":"…", "industry":"…", "event":"…" }
# 2 · score a transaction before it commits
POST https://shield.nemesislabs.xyz/api/v1/omniguard/score       { "function_id":"fn_…", "amount":…, "counterparty_account":"…" }
                         → { "verdict":"block", "overall_score":97, "reasons":[…], "transaction_id":"…" }
# 3 · screen a party (no function needed, screening is free)
POST https://shield.nemesislabs.xyz/api/v1/omniguard/verify      { "check":"sanctions_pep", "subject":"…" }
# 4 · report the outcome + file the STR
POST https://shield.nemesislabs.xyz/api/v1/omniguard/outcome     { "transaction_id":"…", "outcome":"fraud" }
GET  https://shield.nemesislabs.xyz/api/v1/omniguard/reports/{id}/goaml   → goAML STR XML for the NFIU
Create an Omniguard account →About Omniguard
Vibe-code it - Nemesis MCP

Wire all of this from your AI coding agent. Add the Nemesis MCP server to Claude, Cursor, Windsurf or VS Code, then just ask - the agent calls the real tools (create a function, score a transaction, screen a name, protect the app) and ships it for you. Sanctions and PEP screening is free.

1 · Add the server
{
  "mcpServers": {
    "nemesis": {
      "command": "npx",
      "args": ["-y", "@nemesis-shield-autogon/mcp"],
      "env": { "NEMESIS_API_KEY": "nsk_your_key" }
    }
  }
}
Or one command: npx -y @nemesis-shield-autogon/mcp
2 · Ask your agent
Add Nemesis to my exchange: create an Omniguard function for crypto on-ramp and withdrawals, score every movement, screen each counterparty (including the name behind a wallet) against sanctions and PEP lists, and protect the API surface.
Would rather we set it up with you?

Not every team wants to wire this in themselves. Tell us your setup and our engineers will map the integration to your stack and walk you through it end to end, from first call to enforce.

Contact us for hands-on help →

Recommended reading

Positive Security

You Can't Blocklist What Doesn't Exist Yet

AI generates novel attacks faster than any signature can be written. The only defense that scales is learning what your systems are supposed to do, and refusing the rest.

Read · 14 min →

Other industries

Life · General · Health & HMO · Motor

Security for Insurers & Health Insurance

Cut claims fraud and protect member data, on one entity graph, from a single agent.

Read →
Marketplaces · Retail · Travel · Card payments

Security for Online Stores & Payments

Approve more customers, charge back less, and keep the bots and card-testing off your checkout.

Read →
Operators · PSBs · Data Centers

Security for Telecom Operators & Data Centers

Protect the network, the money and your data center, from one agent, with a security line you resell.

Read →