POST to /searches with entity_type (person or company), name, and the filter.types array specifying sanction, pep, adverse-media, and warning
Inspect the hits array in the response; each hit includes a match_status and a score field indicating match confidence
For each hit, iterate over the sources array to identify which list (e.g., OFAC SDN, EU Consolidated, UN Sanctions) triggered the match
Apply a decision workflow: auto-clear low-score hits below your false-positive threshold, escalate medium-score hits for analyst review, and block high-score confirmed matches
Store the search_id and snapshot for audit trail; use it to retrieve updated results if the underlying list changes
Known gotchas
ComplyAdvantage uses a fuzzy matching algorithm; name transliterations and aliases can produce many false positives — tune the fuzziness parameter per language
Adverse media hits are based on web crawl data and may include unverified allegations; treat them as a risk signal, not a confirmed fact
Searching without the birth_year filter on common names will inflate false-positive rates significantly
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp