Today you can ask Claude to plan a trip, draft a contract or refactor a codebase. Until now you could not ask it to fix your chatbot. You had to open the Console, find the conversation, work out why the reply went wrong, edit a FAQ, sync the knowledge base and test again, one click at a time.
With Console MCP in Gydr 4.1, the AI assistant you already use can do that work with you.
Console MCP
Console MCP connects Claude, Codex or any other client that supports the Model Context Protocol to your Gydr Console. Then you ask for what you want in plain words. "Add these twenty FAQs from our returns policy." "Why did the bot tell this visitor we deliver on Sundays?" "Build a widget for our new store and make it match our brand." The assistant does the work and tests the result.
It can add and update FAQs, products and locations, write and test a skill, build a widget and its design, and read conversations to see why a reply came out the way it did. Then it can fix the cause and test again.
It acts as you, with your permissions
The assistant works through the same checks as the Console, using your own role. A teammate with read-only access gets a read-only assistant. It never sees your passwords or your integration secrets.
Anything that changes what your visitors see, or deletes data, asks you first. You stay the one who decides what goes live.
You choose where it can work
Many people work across several Gydr accounts: their own business, a client's, a sister brand's. When you connect an app, you tick the accounts it may work in. Each app keeps its own current account, so switching in one never moves another.
Settings → Connected Apps lists every app, when it was last used and which account it is working in. From there you can change an app's accounts or disconnect it.
On every plan
On the day of launch we made Console MCP available on every Gydr plan, Free included. Each request uses a small number of credits from the account it runs in. If an account runs out of credits or has an overdue payment, its requests are refused before anything is sent.
Why we built it
We believe the way people operate software is changing. For decades, every product had to be learned screen by screen. More and more, people will describe what they want and let an assistant they trust carry it out, across every tool they use.
For that to work, two things have to be true. The software has to be reachable by AI. And it has to stay safe when it is: the same permissions, the same confirmations, the same audit trail as a person clicking through it. Console MCP is our attempt to do both.
Frequently asked questions
What is Console MCP?
A connection that lets an MCP client such as Claude or Codex operate your Gydr Console as you, with your own Console role and permissions.
Which Gydr plans include Console MCP?
Every plan, Free included, since Gydr 4.1.1. Each request uses a small number of credits from the account it runs in.
Can the AI assistant change my live chatbot without asking?
No. Anything that changes what visitors see, or deletes data, asks you first.
Setup takes a few minutes and is described in the Console MCP documentation. The full list of changes is in the 4.1.0 release notes.