AMLBot - Crypto Risk Screening

Available To

Free for the following exchange plans:

  • Cloud Plans:

    • Enterprise

  • On-Premise Plans:

    • Enterprise Unlimited

What Is It?

The AMLBot Plugin is a crypto AML and KYT (Know Your Transaction) screening solution integrated into HollaEx exchanges. Powered by AMLBotarrow-up-right, it automatically reviews on-hold deposits and withdrawals and decides whether they are safe to release or should be kept for manual review.

For each pending transaction, the plugin queries the AMLBot KYT API with the on-chain transaction hash (deposits) or destination address (withdrawals). AMLBot returns an overall risk score (0–100%), a granular breakdown of risk signals (sanctions, dark markets, scams, stolen coins, ransom, terrorism financing, mixers, etc.), and a blacklist flag.

Transactions whose risk profile exceeds your configured thresholds stay on hold, and the audit team is alerted by email. Transactions are auto-released only after AMLBot returns a completed/successful result below the configured thresholds.

Who Needs It?

This plugin is essential for any HollaEx exchange operator that:

  • Needs to comply with AML, sanctions, and Travel Rule regulations.

  • Wants to automatically catch deposits/withdrawals tied to scams, dark markets, sanctioned actors, or stolen funds before settlement.

  • Prefers a percentage-based, tunable risk engine instead of a binary blocklist.

  • Wants to reduce the operational load on compliance reviewers by auto-releasing low-risk traffic.

How to Use It?

You can simply install the plugin from the Plugins section inside the Operator Control. After installation, configure the plugin meta with your AMLBot credentials and risk thresholds.

1. Get your AMLBot API credentials

  1. Open your Profile → API Keys.

  2. Generate an accessId and accessKey pair and store them somewhere secure. They are used together to compute the MD5 token sent on every API call.

2. Enable manual review of deposits and withdrawals

The plugin only acts on transactions that have been placed on hold. Disable the auto-processing flag for each flow you want AMLBot to screen:

  • kit.auto_deposit.active = false for deposit screening.

  • kit.auto_withdrawal.active = false for withdrawal screening.

If both are enabled, the plugin will log that there is nothing to process and exit each cycle.

The plugin does this for you automatically on first install. When the plugin starts for the first time it will switch both auto_deposit and auto_withdrawal off in your kit configuration and send a one-time alert email to the audit recipient (or to the address configured in alert_email) explaining what changed and what it means for transactions. After that, the change is recorded in Redis and will not be reapplied — you remain free to re-enable either toggle at any time from Operator Control → General → Security, but doing so will silently disable AMLBot screening for that flow because there will be no on-hold queue for the plugin to read.

3. Configure the plugin

Open the plugin in the Operator Control and set the following fields:

Field
Required
Default
Description

api_url

no

https://extramlbot.com

AMLBot KYT check endpoint URL. The plugin posts directly to this URL and falls back to /check/ only if the endpoint is unavailable.

access_id

yes

AMLBot accessId.

access_key

yes

AMLBot accessKey. Used together with access_id to compute the MD5 token for each request.

min_usdt

no

100

Minimum USDT-equivalent value before AMLBot is queried. Anything below this is auto-released without an external check.

risk_threshold_percent

no

50

Overall risk score percentage (0–100) at or above which a transaction is held.

high_risk_signal_threshold_percent

no

1.5

Combined percent of high-risk signal categories (sanctions, dark_market, scam, stolen_coins, terrorism_financing, etc.) at or above which a transaction is held, regardless of overall score. AMLBot signal fractions are converted to percent before comparison.

risk_threshold

no

HIGH

Discrete severity bucket at or above which a transaction is held. Allowed: CRITICAL, HIGH, MEDIUM, LOW.

block_on_blacklist

no

true

Hold any transaction with an AMLBot blacklist flag, regardless of risk percentage.

check_withdrawals

no

true

Set to false for deposits-only mode.

flow

no

fast

AMLBot analysis flow. fast returns quicker; advanced returns the full report.

cache_ttl_hours

no

48

How long to cache risk results per address/transaction.

request_timeout_ms

no

15000

Per-request timeout for AMLBot calls.

alert_email

no

(audit email from kit secrets)

Optional override for the recipient of risk-hold alert emails.

4. (Optional) Trigger a manual run

The plugin runs automatically every 60 seconds. To trigger an immediate cycle, send an authenticated request from an admin account:

What gets blocked

A pending deposit or withdrawal is kept on hold when any of the following is true:

  • block_on_blacklist is on AND the address has an AMLBot blacklist flag, OR

  • The overall risk score percentage is greater than or equal to risk_threshold_percent, OR

  • The sum of high-risk signal percentages is greater than or equal to high_risk_signal_threshold_percent, OR

  • The discrete severity bucket meets or exceeds risk_threshold.

Pending, failed, or otherwise incomplete AMLBot API statuses are kept on hold and retried in a later cycle; they do not trigger a risk alert because the risk calculation is not ready. Otherwise, the transaction is automatically released. When a transaction is held due to risk, an alert email is sent to the configured recipient (or the kit's audit email) with the severity, percentages, blacklist flag, and the AMLBot UID for follow-up.


Benefits for HollaEx Operators

The AMLBot plugin gives exchange operators continuous, automated AML coverage without growing the compliance team. It blocks high-risk deposits and withdrawals before settlement, auto-releases the long tail of clean traffic, and gives you fine-grained control over how strict the engine should be — by overall score, by signal category, or by simple severity buckets. The result is stronger AML posture, fewer manual reviews, and faster customer experience for legitimate users.

Last updated