MCP server

Custom Webhook MCP for Your Agents

The Custom Webhook MCP is a first-party connector that turns any HTTP endpoint into a tool your AIML.chat agents can call — no MCP server to build — with human write-confirm before any action that changes data. It installs from the MCP marketplace and is the fastest way to wire your own system in.

The Custom Webhook MCP is a first-party connector that turns any HTTP endpoint into a tool your AIML.chat agents can call — no MCP server to build — with human write-confirm before any action that changes data. It installs from the MCP marketplace and is the fastest way to wire your own system in.

What you can do with a custom webhook

When no marketplace connector exists for your system, you don't have to build a full MCP server. The Custom Webhook option lets you point an agent at one of your own HTTP endpoints — an internal lookup, a provisioning action, a notification hook — and describe what it does and what inputs it takes. The agent then calls it as a tool when relevant. Read-style endpoints (a lookup) can answer with live data; action endpoints (something that changes state) run behind AIML.chat's write-confirm flow, so the agent proposes the call and a person approves before your endpoint is hit. It's the simplest path from "we have an internal API" to "our agent can use it."

How it works

You provide the endpoint URL, method, and a description of the inputs and what the call does, plus auth (none, a bearer token, or a custom header) — all stored encrypted. You mark whether the endpoint changes data; if it does, every call requires approval. Because you define exactly one endpoint per tool and scope what the agent may send, you expose precisely what you intend and nothing more. For richer, multi-tool integrations, build a full MCP server using the agent SDK template instead.

Connect it in three steps

  1. 1In AIML.chat, open MCP → Marketplace → Custom Webhook and click Add.
  2. 2Enter your endpoint URL, method, input description, and auth — and mark whether it changes data.
  3. 3Attach the tool to an agent and test: trigger the call and, for an action endpoint, approve the write.

Most teams start with a read-style lookup, then add an action endpoint once they're comfortable with the approval flow. You start on a real free tier and scale as usage grows — see pricing for the agent and message limits, and the MCP docs for the contract. When you're ready, start free.

Do I need to build an MCP server?+

No. The Custom Webhook option maps a single HTTP endpoint to a tool with no server to build; for multi-tool integrations, use the SDK template.

Will it call action endpoints on its own?+

No. If you mark an endpoint as changing data, every call requires human approval first.

Is there a free plan?+

Yes — start on the free tier. See pricing for higher agent and message limits.

Add AI chat to your site

Free plan · No credit card · 5 minutes