Make MCP Feature Extention

Questions
1

I have been using the Make MCP Feature extensively to run actions within Clause as part of my Human In the Loop experiment.

The opportunities using the make mcp client are huge:

  1. I dont have to have lots of mcp connections
  2. I can dsign tooling to support how I work
  3. My team can access tooling and intelligence from within their LLm’s

BUT

And here are the limitations.

The MCP token gives access to all on-demand scenarios. This becomes slightly problematic when you are also building agents, additionally the token seems to pull as OD scenarios access teams, so that route is not a solution.

What we should be able to do is have 1) MCP ‘tokens’ restricted to teams (if that is part of your subscription), 2) Be able to ‘assign’ OD scenarios to each MCP token.

I have spoken to some of my clients, and here is why this will be a helpful feature extension.

  1. I can create a team for each client.
  2. Within that team, set up MCP, which are
  • Company-wide
  • Individual or operational team specific

This way we could tune MCP actions for example only select people could have an MCP with an action to update YouTube descriptions, or only select people could have an MCP which updates blog images from within Claude (or whichever LLM you are working in)

The current connection is great as a PoC and I have already shown clients what I am able to do with it and they are keen. But with the current execution of the MCP client its all or nothing, which is a missed opportunity not just for Make.com but for us AI Transformation experts.

The whole MCP connection should be viewed as a product within Make - I think it warrants it - Make could even charge for it as an add-on feature/plan so you would have an add-on sub and then the related credits. I would subscribe to that :slight_smile: .

The alternative for me is to create custom MCPs that use the make API - which is a ton of work - and I dont want to waste time going down this route.