Use the embedded agent
Nothing else to connect. Log in, start chatting and upload your product files. The agent interviews you and builds the draft inside the platform.
AI + MCPChat your product
Chat your product
into existence
Log in and work with the AI agent embedded inside InoatriX. Give it your specification files, answer its questions and let it configure the product with you. Switch to the no-code studio at any moment—or bring your own AI agent through MCP.
CHOOSE YOUR AI
Built in when you want it. Open when you need it.
Use the agent already inside the platform or connect your preferred one. Either way, your team can take over visually at any point.
Bring your own agent through MCP
Connect any assistant that speaks MCP — the open standard for linking AI to outside tools — and create governed drafts with the AI tooling your team already uses.
How to connectMove between AI and no-code
Both agent paths and the visual studio operate on the same product. Switch tools without exporting files, recreating work or losing context.
THE EMBEDDED FLOW
From source files to a reviewable product, in one conversation.
The agent does not just ingest a specification. It works with you to make the product complete, then lets you continue in the interface whenever that is faster.
Log in and give the agent your specification files
Start a conversation inside InoatriX and upload the documents and spreadsheets you already use to define the product. The agent reads them in the context of the platform's real configuration capabilities.
The agent interviews you about what is missing
It identifies gaps and ambiguities, asks focused follow-up questions and confirms the decisions needed for risk objects, coverages, plans, rules, lifecycle and pricing.
It validates and configures a DRAFT
The agreed specification is checked before it is written. The agent configures the product in the platform and keeps you informed as the draft takes shape.
Switch seamlessly between chat and no-code
Open the same draft in the Product Configurator, inspect or change any setting yourself, and return to the embedded agent whenever you want. Your team remains in control of publishing.
THE MCP GUARDRAILS
Your connected agents get a deliberately bounded route into the platform.
The interesting question about AI configuration is not what it can build. It is what it cannot reach.
It cannot publish
Publishing is not something a connected assistant can do — the ability is absent from the connection, not merely discouraged. Going live stays a human decision.
It deletes only drafts
An assistant can tidy up drafts it created and nothing else. A product that is live, or that someone else is working on, is out of its reach entirely.
Nothing is a black box
Every value the agent set is a normal field in the configurator. There is no AI-only state, and no setting your team cannot see, change or override.
Partial builds are reported, not hidden
If a build step fails, the partial draft is kept and the failure is reported—so you inspect what exists rather than guess what happened.
Scoped credentials on both sides
Each connected assistant gets its own private key, and acts under an identity of its own. Every draft carries the name of whoever created it, so it is always clear which assistant did what.
An honest deliverable
A configured draft is structurally complete but not pretending to be underwritten. It is a starting point for review—which is exactly what it is presented as.
CONNECT YOUR OWN
Already working with an AI assistant? Point it at InoatriX.
Any MCP-capable assistant connects — ChatGPT and Claude are simply the two we publish ready-made settings for. It takes about fifteen minutes and no code: a key, an address, and a restart.
A step-by-step setup guide
Walks through every step in plain language: the key we issue you, pasting the settings, checking the connection worked, and what to do if it does not.
THE POINT
One product. A continuous conversation and workspace.
AI speed. Human continuity.
Upload what you know, let the agent ask about what you do not, and open the visual studio whenever direct configuration is the better tool. Every route leads to the same governed draft, and the decision to go live stays with the people accountable for it.
SEE IT HANDS ON
Book a demo