AML checks inside your AI assistant
Connect your Kytme account to Claude, ChatGPT or any other MCP-enabled assistant and run AML checks by simply asking. Risk scores, sources of funds, reports, balance and address monitoring — without leaving the conversation. Your first connection comes with a free check.
Connector address:https://mcp.kytme.com/mcp
Works with the assistants you already use
- Claude
- ChatGPT
- Grok
- Gemini
- any MCP client
What the connector can do
The full Kytme toolkit, exposed to your assistant as safe, permission-scoped tools.
Run AML checks by asking
Ask your assistant to check an address or a transaction. It quotes the exact price first and runs the check only after you confirm — no silent spending.
Balance, history and reports
See how many checks you have left, search your check history by address, network, risk or date, and open the full web report of any result.
Address monitoring
Put addresses on scheduled re-checks, tune the frequency and alerts, and review how the risk of an address changed over time — all from the chat.
Your risk models
View and adjust your custom risk models per provider: which risk categories to ignore and from which level to raise an alert. Sanctions categories always count.
What it looks like
Ask in plain language — the assistant quotes the price, waits for your yes, and does the rest.
Check this BTC address: bc1q0kw…qp5cr4xt0
This is a paid check: 100 units (1 check), balance after — 400 units. Run it?
Yes, go ahead
Done — low risk, 12/100. Trusted sources: 96%. Full report: kytme.com/report/…
Watch this address for me, re-check it weekly
Weekly monitoring via BitOK: 100 units per check, about 400 units a month. Set it up?
Yes, with email alerts
Monitoring started: weekly checks, alerts to your account email. Risk history will build up in My Addresses.
How to connect
- Step 1
Create a Kytme account — or sign in to the one you already have.
- Step 2
Add a custom connector in your assistant and paste the address https://mcp.kytme.com/mcp. In Claude: Settings → Connectors; in ChatGPT: Settings → Apps & Connectors (Developer mode).
- Step 3
Sign in on the Kytme consent screen and choose what the assistant may access. Your first connection grants a free check — then just ask for your first AML check.
You stay in control
The connector uses OAuth: you sign in on kytme.com and see exactly which permissions the assistant gets — reading checks, spending checks, managing risk models or monitored addresses. Paid actions always show the price and wait for your confirmation, alert recipients cannot be changed from the chat, and you can revoke access at any moment on the connected apps page. Checks run on the same risk engine as the dashboard and API, and stay in your account history.
FAQ
- Which assistants are supported?
- Claude (web, desktop and Claude Code), ChatGPT (custom connectors in Developer mode), Grok, Gemini — and any other client that speaks MCP, including Cursor and VS Code. The connector is a standard MCP server.
- Can the assistant spend my balance without asking?
- No. A paid check is a two-step action: the assistant first receives a signed quote with the exact price and has to show it to you; the check runs only after your confirmation, and the quote works exactly once.
- Is the connector free?
- Connecting is free, and the first connection comes with a bonus check. Checks themselves are paid from your Kytme balance at the same per-check price as in the dashboard.
- How do I disconnect an assistant?
- Open the connected apps page in your Kytme account and revoke the access — the assistant loses its tokens within seconds. Changing your password revokes access too.
- Is this the same as the Kytme API?
- No. The API is a separate high-volume integration channel with its own keys. The connector is your personal account in chat form: it acts on your behalf, under the permissions you granted, with per-hour limits.