DART
Digital Asset Risk Targeting
Know where the money comes from.
Keep bad funds out.
DART traces every fund-flow chain across labeled addresses and turns any address or transaction into an explainable 0–100 risk score — you define which chains count as risky, and how many points they cost.
How a screening runs
One screening is completed by two independent rulesets: the path-matching ruleset decides “what counts as a violation”, the scoring ruleset decides “how severe it is and what to do”.
- INPUT
Address / TX hash
Submit an address (KYA) or a transaction (KYT). Chain and token are auto-detected.
- STEP 01
Path investigation
Expand the fund-flow graph hop by hop — upstream and downstream — with a label on every counterparty.
- STEP 02
Path-matching ruleset
Every path is matched against 72 rules: which fund-flow chains count as risky — and which rule was violated.
- STEP 03
Scoring ruleset
Hits are converted into a 0–100 score by direction, hop distance, severity and share of funds.
- OUTPUT
Score + disposition
Pass · Review · Enhanced Due Diligence · Block — with the exact rules behind the number.
Two rulesets, one screening
You bring the definitions: which fund-flow chains are risky, and how many points each should cost. DART ships 72 built-in rules per ruleset, fully customizable.
“Is this fund-flow chain a violation?”
- 72 rules per ruleset across 9 risk categories — sanctions, terrorism financing, mixers, hacks, darknet and more.
- Three rule shapes: self-hit (0 hop) · direct exposure (1 hop) · layered exposure (2–5 hops, with rising materiality thresholds).
- Input: fund-flow graph + node labels. Output: a hit list — rule, severity, hop, amount.
- Copy a built-in ruleset and edit any threshold or condition.
“How severe is it — and what should we do?”
- A contribution matrix over direction × hop distance × hit severity — inflow vs outflow weighted differently.
- Cell score = base × severity weight × share of the address's total on-chain flow (not the sub-graph).
- Total capped at 100, mapped to four dispositions: Pass · Review · EDD · Block.
- Self-hit rules override everything: score 100 × severity weight → Block.
The path-matching ruleset only answers yes/no — it never produces a score. The score comes entirely from the scoring ruleset. So the same hit list scores differently under a different scoring ruleset, while the hits themselves never change.
One score, four dispositions
Every screening ends in a 0–100 score and one of four dispositions — with the exact rules that produced it, ready for your audit trail.
No risk exposure found that needs action.
Risk linkage exists — manual review advised.
Significant risk — enhanced due diligence required.
Extreme risk — refuse the transaction.
One engine, four capabilities
Everything a compliance team needs to screen crypto flows — built on our own data, our own labels, our own rules.
Address Screening
Screen any address for direct hits and indirect exposure across 0–5 hops — sanctions, terrorism financing, mixers, hacks, darknet and more. Materiality thresholds are fully configurable.
Transaction Screening
Direction-aware screening for deposits and withdrawals. Trace the funds behind each transfer back to their source, and block or escalate in real time.
Sanctions Screening
Self-built sanctions database covering OFAC, UN, EU, UK and more — entities, individuals, companies and addresses, searchable in under a second.
Intelligence & Labels
Continuously updated labels for exchanges, DeFi protocols, mixers, hacks and darknet entities — verified by a dedicated intelligence team.
Learn the product in an afternoon
Screenshots, console walkthrough, API samples — everything about the two rulesets, from your first screening to your own scoring matrix.
DART Product Manual
The two rulesets explained end to end: how a screening runs, all 72 rule categories, the scoring matrix and formulas, console + API usage with screenshots.
Read the manual →API Reference
KYA / KYT endpoints, ruleset_id and scoring_ruleset_id parameters, sync and async modes, and the result schema — testable right in the browser.
Open API referenceAMLClaw — open-source app
A ready-to-run compliance desk built on DART's API — screening, address/TX monitoring and evidence reports — and a working example of how to call the API well.
View on GitHubQuestions we get asked
Latency, coverage, accuracy and explainability — the four things every compliance team checks before integrating.
Near zero.
DART runs a database-first, RPC-fallback pipeline: a query first hits our indexed on-chain database, and when the transaction or address is not indexed yet, the system falls back to reading it from a node over RPC. So even a transaction that has just landed on-chain is screened the moment you query it — there is no batch job and no sync window to wait for.