November 21, 2026 · The CompliAPI team
Real-Time vs Batch Sanctions Screening
Real-time sanctions screening checks a party at the moment of a decision; batch screening re-checks the parties you already hold. What each catches, what each misses, and how to run both.

Real-time sanctions screening checks a party at the moment of a decision: one request, one answer, before the account opens or the payment leaves. Batch screening re-checks the parties you already hold, on a schedule or when a list changes. They answer different questions, and a sanctions program that only runs one of them has a gap.
This post sets out what each mode is for, where each one fails, and how the two fit together with a third mode, ongoing monitoring. The examples use CompliAPI's endpoints, but the reasoning applies to any screening tool.
What "real-time" has to mean
The phrase covers two separate properties, and a tool can have one without the other.
The check runs at decision time. The screening call sits inline in the flow, and the product waits for the result before it proceeds. A withdrawal is not sent, an account is not activated, a listing does not go live until the answer is back.
The data is current. The result reflects the list as the authority last published it, not as it stood at the last nightly load. OFAC publishes changes to the SDN list at any time of day, and a designation is effective when it is published.
A check that is fast but runs against stale data returns a confident "clear" for a party designated that morning. When you evaluate a provider, ask how often each list is refreshed and whether the response tells you which version it checked.
CompliAPI's answer: the OFAC SDN and Consolidated lists are checked for a new Treasury publication every 15 minutes and ingested when one appears. The stablecoin issuer freeze lists are crawled on the same 15-minute cycle. The other sanctions lists (UN, EU, UK, Switzerland, Canada, Australia, New Zealand, France, Japan) refresh daily. Every screening response includes list_versions, the last refresh time of each list checked and, for the OFAC lists, Treasury's publication id. The schedule for every list is on the data sources pages. We publish no response-time figure and no uptime commitment, so measure latency from your own infrastructure before you put a check inline.
Real-time screening: one decision, one request
A real-time check is a single authenticated GET request:
curl https://api.compliapi.com/api/v1/screen/crypto/0x098B716B8Aaf21512996dC57EB0615e2383E2f96 \
-H "Authorization: Bearer $COMPLIAPI_TOKEN"
The response says whether the value is sanctioned (a sanctions-list hit) or flagged (any list hit, including crime and risk lists), names the list behind each match, and links the official record. The same shape applies to emails, websites, government IDs, vessels, aircraft and countries. Name screening is fuzzy and returns scored candidates.
Where it belongs:
- Onboarding, before the relationship starts.
- Before value moves: a withdrawal address, a payout recipient, a counterparty on an invoice.
- Per session, for location rules: the country behind an IP address and whether a VPN is hiding it.
What it cannot do is look backwards. It says nothing about the customer you cleared last quarter.
Batch screening: the parties you already hold
Batch screening takes a set of stored parties and checks all of them against the current lists. It exists because lists change after you onboard someone.
CompliAPI's batch endpoint screens up to 50 subjects in one request, with types mixed freely:
curl -X POST https://api.compliapi.com/api/v1/screen/batch \
-H "Authorization: Bearer $COMPLIAPI_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"items": [
{"type": "name", "value": "Jane Example", "birth_date": "1980", "ref": "cust-1042"},
{"type": "crypto", "value": "0x098B716B8Aaf21512996dC57EB0615e2383E2f96", "ref": "wallet-77"}
]
}'
Each item is screened exactly as its own endpoint would screen it and comes back in request order with your ref. Each item counts as one request against your monthly quota, and an item whose input is rejected is not counted. A customer base larger than 50 is a loop over pages of 50.
Where it belongs:
- Scheduled re-screening of the whole book.
- Back-book checks when you add a list to your policy or migrate from another provider.
- Screening everything known about one customer in a single call: name, email, ID and wallet together.
Its weakness is the gap between runs. A nightly batch can leave a newly designated party unnoticed for most of a day.
Ongoing monitoring: alerts when a list changes
The third mode inverts the direction. Instead of asking the list about your parties, you register the identifiers you hold and are told when a list changes for one of them.
In CompliAPI this is sanctions list monitoring, on the paid plans. You add wallet addresses, emails, websites or government IDs; when one is listed, delisted or relisted, a signed webhook or an email goes out within a few minutes of the change being ingested. Two limits matter. Matching is exact, so monitoring covers identifiers and not names. And an alert fires on a change, so a value that was already listed when you added it does not alert for that listing: screen it first.
Side by side
| Real-time | Batch | Monitoring | |
|---|---|---|---|
| Question answered | Is this party listed now? | Are any of the parties I hold listed now? | Has a list changed for a party I hold? |
| Trigger | A decision in your product | A schedule, or your own job | A list update |
| Catches | A listed party at the door | A party listed since onboarding, at the next run | A party listed since onboarding, within minutes of ingestion |
| Misses | Later designations | Anything between runs | Names; values listed before they were added |
| CompliAPI surface | GET /screen/* | POST /screen/batch | Monitored values with webhook or email alerts |
How they fit together
A workable default for a product that moves money:
- Screen in real time at every decision point. This is the control the sanctions prohibition maps to most directly: do not deal with a listed party.
- Monitor the identifiers you store, so a later designation reaches you without waiting for a job to run.
- Batch re-screen names and anything monitoring does not cover on the schedule your policy sets, and keep the
list_versionsfrom each run as evidence of what was checked.
How often to run step 3, and what happens on a match, are decisions for your compliance team. Teams in regulated finance will find the same three modes on the sanctions screening for banks page, applied to onboarding, payments and online-banking sessions.
A screen is a fact about a list at a point in time. It is not a compliance determination, and no screening mode removes the need for a person to review a match.
Frequently asked questions
What is real-time sanctions screening?
Real-time sanctions screening is checking a name, wallet address or other identifier against the sanctions lists at the moment a decision is made, before an account opens or a payment leaves, and getting the answer back in the same request. It has two parts that are easy to conflate: the check runs at the time of the decision, and the list data behind it is current. A check that runs instantly against a week-old copy of the list is not real-time in the sense that matters.
Is batch screening still needed if every transaction is screened in real time?
Yes. A real-time check tells you a party was clear when you checked. Lists change afterwards: a customer onboarded in March can be designated in June without sending a single transaction. Re-screening the parties you hold, on a schedule or when a list changes, is what catches that. The two are complements, and most sanctions programs run both.
How often should a customer base be re-screened?
No regulation sets one interval; it is a risk-based decision your compliance team makes and documents. The data side sets the useful floor: there is nothing new to find between two publications of a list. Many teams re-screen when a list they rely on changes and run a full sweep on a fixed schedule as a backstop. Monitoring that alerts on list changes replaces the first of those for the identifiers it covers.
How fresh is the list data behind a real-time check?
It depends on the provider, and you should ask for the number. CompliAPI checks for a new OFAC publication every 15 minutes and ingests the SDN and Consolidated lists when one appears; the stablecoin freeze lists are crawled every 15 minutes; the other sanctions lists refresh daily. Every screening response carries list_versions, the last refresh time of each list it was checked against, so the freshness of a given result is on the result itself.