Agent integrations
Connect FinCrime Desk to your AI assistant
The live role index, the enforcement record, and your own employer watchlist — queryable from Claude, ChatGPT, or any client that speaks the Model Context Protocol. Ask your assistant what's open in Singapore sanctions this week, or have it add an employer to your watchlist without leaving the chat.
01 · Endpoint
https://fincrimedesk.com/mcpRemote MCP over HTTP. Authorization: OAuth 2.1 — you sign in with your FinCrime Desk account and tools respect your permissions.
02 · Tools
search_jobsSearch the live index of open financial crime compliance roles by keyword, discipline, location, and seniority.search_companiesLook up employer profiles — sector, headquarters, and current hiring activity.list_enforcement_actionsRead the regulatory enforcement record: regulator, respondent, penalty, and date.list_watched_companiesList the employers on your own watchlist.watch_companyAdd an employer to your watchlist.unwatch_companyRemove an employer from your watchlist.
Read tools return the same public data as the site; watchlist tools act only on your own account.
03 · Connect
Claude
- 01Open Settings → Connectors and choose “Add custom connector”.
- 02Name it FinCrime Desk and paste the endpoint URL above.
- 03Approve the sign-in prompt with your FinCrime Desk account.
- 04Enable the connector from the chat composer and ask away.
ChatGPT
- 01Open Settings → Connectors and enable developer mode if prompted.
- 02Create a new connector and paste the endpoint URL above.
- 03Complete the sign-in prompt with your FinCrime Desk account.
- 04Turn the connector on in the composer and ask away.
Any other MCP client
Register a remote MCP server over streamable HTTP pointing at the endpoint above. The server advertises its own OAuth metadata, so a compliant client will register itself and walk you through sign-in. After we ship changes, refresh or reconnect the server so your assistant picks up the current tool list.
04 · What this is not
This is not a bulk data feed. Tool responses are page-sized and rate-limited, and the same limits and permissions that apply to you on the site apply to your assistant. Pay figures remain self-reported and unverified; enforcement records are reproduced from regulator publications. See Methodology for how each dataset is built.