Use Make with AI via MCP
If your processes already run in Make, nothing has to be rebuilt for AI: the official Make MCP server turns existing scenarios into tools for agents. More than 3,000 apps with around 30,000 actions become reachable for CompanyGPT – without adding a single new interface.
What is Make?
Make is a visual automation platform. Processes are modeled as scenarios in which modules from different applications are linked, data is transformed and conditions are expressed. The catalogue spans more than 3,000 apps with roughly 30,000 actions, complemented by generic HTTP modules for services without a ready-made connector.
What makes Make interesting in an AI context is the direction of the connection: the platform is not opened up as a whole – the scenarios you have already defined become callable tools. What an agent can trigger therefore follows from what your team has approved and tested anyway.
Connection at a Glance
- MCP server
- Official (Make), remote
- Endpoint
- mcp.make.com or token-based server URL
- Transport
- Streamable HTTP
- Authentication
- OAuth or token
- Operating model
- SaaS
- Reach
- More than 3,000 apps, around 30,000 actions
How to Connect Make to CompanyGPT
Make provides an official MCP server. The connection runs through mcp.make.com with OAuth, or through a token-based server URL. Streamable HTTP serves as transport, and CompanyGPT connects to the endpoint directly.
The heart of the integration: your scenarios become tools. An agent sees the scenarios released for MCP access and can run them with the parameters they define. The scope therefore stays exactly as wide as you modeled it – an agent gains no general permissions in the connected target systems.
In CompanyGPT your administrators then assign the integration via roles and groups. In practice this gives individual departments precisely the automations that fit their work.
Typical Use Cases
How AI agents in CompanyGPT work with Make
Start existing automations from chat
A scenario already in production – creating a draft quote, for instance – is triggered right in the assistant, without opening the Make interface.
Reach niche applications
Systems without an MCP server of their own are connected through a scenario and become available to the agent as a clearly bounded tool.
Hand data between systems
The agent passes structured information to a scenario that distributes it to several target systems and reports the status back.
Frequently Asked Questions
Make and AI – the most important answers
Does Make have an official MCP server?
Yes. Make provides an official MCP server, connected through mcp.make.com via OAuth or through a token-based server URL.
What exactly does the AI agent see?
The scenarios released for MCP access. They appear as tools with their defined parameters – not as general access to your Make organization.
Do we have to build new scenarios?
Not necessarily. Existing scenarios can be used as long as their inputs and outputs are suitable for being called by an agent. Small adjustments to the parameters are often enough.
How is an agent kept from triggering too much?
Through two layers: Make only exposes the scenarios you released, and in CompanyGPT roles and groups control which users may use the integration at all.
How does this compare to running our own automation platform?
Make is a hosted service, so execution happens there. If automations should deliberately run inside your own environment, a self-operated platform is the better foundation. We work that assessment through with you.
Make and Your AI – GDPR-Compliant with CompanyGPT
CompanyGPT is the enterprise AI assistant in your own cloud tenant: all relevant AI models, your company data and integrations like Make – centrally managed, with roles and permissions. We set up the connection together with you.
Make is a trademark of its respective owner (Make). Its mention describes compatibility and does not imply any partnership or endorsement. Availability and scope of the connection depend on the respective vendor.