The assistant your team works in, Claude, ChatGPT or Gemini, can now call complead's tools directly, with your account. We didn't add a chat window to complead. The connection runs over MCP, the Model Context Protocol, an open source standard for connecting AI applications to outside systems. You ask in plain English, and the assistant picks the tool, runs it and shows you the result. The analyst decides what happens to the case.
In Claude's directory, the connector appears as Sanction Scanner. That was the company's name until 29 September 2026, when complead.ai launched and Sanction Scanner became complead. Sanction Scanner, the screening product, is part of complead, and the Claude listing still carries that product name. The tools reach the same screening, cases and monitoring that run on the complead platform.
Where can you find complead today?
As of 18 November 2026, there are four ways in: three directory listings and one URL.
- Claude. complead is available in Claude's connector directory as Sanction Scanner, listed there since June 2026.
- ChatGPT. complead is listed in ChatGPT's directory.
- Gemini. complead is listed in Gemini's directory.
- Any MCP client. Add complead as a custom MCP server with the URL
https://api.sanctionscanner.com/mcp, the same URL the Claude listing uses.
On every route you sign in with your complead account. Authentication uses your organization's existing API credentials, so there's no separate account to create. That account also decides which products the tools can reach.
What can you ask complead from Claude, ChatGPT or Gemini?
Picture an analyst with four requests on a Tuesday morning. The company and the person in them aren't real.
- "Screen Example Freight Ltd against sanctions and PEP lists and summarize any matches." The assistant calls
scan_by_name, thenget_scan_summary. - "Run an adverse media search on John Doe and show the financial crime articles." It calls
scan_adverse_mediaandget_adverse_media_detail. - "This match is a false positive. Mark it, add a memo with the reason and set the risk level to low." That's
set_match_status,add_case_memoandset_risk_level: three changes you approve. - "Show this week's monitoring alerts and the open workload per analyst." It reads
list_monitoring_alertsandget_organization_workload.
Behind those four requests sit 30 tools. Twenty read: scan summaries, customers and their risk scores, company profiles, adverse media, alerts, monitoring history and compliance statistics. Two start a new scan, scan_by_name and scan_adverse_media, and the other eight write.
The name screening behind scan_by_name is complead's client screening and ongoing monitoring. It uses fuzzy matching, and its coverage stands at 3,500+ sanctions, PEP and watchlists from 220+ jurisdictions, refreshed every 15 minutes.
Other tool groups need their own product. Customer risk scores need the customer risk assessment product, and company search and profiles need the KYB product.
What can the assistant change, and what stays with the analyst?
Eight tools write to complead. Each one changes a case or a monitoring setting.
set_match_statussets the status of a match, for example false positive.add_case_memoadds a note to the case, such as the reason for a decision.set_risk_levelsets the risk level.add_tagtags a case.assign_scan_to_userhands a case to a named colleague.close_casecloses the case.enable_monitoringstarts ongoing monitoring for a name.set_monitoring_periodsets the monitoring period for that name.
The assistant proposes the call and you approve the change in the assistant. That approval step belongs to the assistant, so each provider sets how it works. None of the 30 tools deletes a case or changes a user's permissions.
Your screening settings, review steps and reports stay as they were. The analyst who approves a change owns the decision.
Is every action logged?
Yes. Every action runs inside your organization's complead account, and complead records it in that account's activity log, whichever assistant made the call. Screening searches also enter complead's search audit trail. It shows who ran each search, what they searched for, when, and what came back, including searches that found no match.
The log covers what happens in complead. Results the assistant reads become part of your chat with that provider, so the provider's own data terms apply there. Read those terms, and your provider's admin settings, before you connect.
How do you connect complead to your assistant?
| Route | Where to find it | Listed name | What you need |
|---|---|---|---|
| Claude | Claude's connector directory | Sanction Scanner | Your complead account to sign in |
| ChatGPT | ChatGPT's directory | As shown in the directory | Your complead account to sign in |
| Gemini | Gemini's directory | As shown in the directory | Your complead account to sign in |
| Any MCP client | Add a custom MCP server with https://api.sanctionscanner.com/mcp | You choose it | A client that supports remote MCP servers with OAuth sign in |
In the three directories you find the listing and sign in with your account. The custom route signs in with OAuth and PKCE, and the server supports dynamic client registration, so a client can register itself before you sign in.
Each provider sets its own rules for custom servers on plans, regions and write access. Claude allows custom connectors on its free and paid plans. On Team and Enterprise plans an owner or a permitted role adds them for the organization before members connect. ChatGPT and Gemini each publish their own conditions for custom MCP servers in ChatGPT and for adding an MCP server to Gemini.
“Our interface was never the point. The ground under it is: the lists, the cases, the record of who decided what. If your analysts already work in Claude, ChatGPT or Gemini, that's where the ground has to be. The decision stays where it was, with the analyst who approves the change. And the log shows who did.”
Fatih Coşkun, Founder and CEO, November 2026
Which route should your team pick?
It depends on who sets it up. When analysts connect it themselves, start with the directory listing for the assistant they use. Your provider's plan and admin settings decide who can add it.
If an engineer runs your MCP setup, or your assistant isn't in those directories, add the custom MCP URL instead.