Feedmind
Capability · AI tool access

Your Feedmind data, readable by your own AI tools.

Generate a read-only token and the AI tool you already use can read your Feedmind account directly: your score, your four dimensions, which products are dragging them down, and what changed this week. We sell AI readability. Our own data being agent-readable is the same argument, applied to us. On every plan, including Free.

The problem

The dashboard is where we put the answers. It is not where you spend your day.

A merchant checking on a catalog opens a tool, asks a question in words, and expects the answer in the same place. Feedmind runs a read-only server for exactly that: your assistant asks it, gets your numbers, and renders them however it renders things. Nothing has to be exported, and nothing can be changed.

The answers come back with the same numbers the dashboard shows, and the same sentences about what they mean.

Ask your AI tool
What is my Feedmind score and which dimension is worst?
Which of my products did Google disapprove, and why?
What changed in my catalog this week?
Answered from your account · read-only
How it works

One token, read-only, revocable.

01

Create a token

Under AI access in your settings. It is shown once, stored as a hash, and named by you so you can tell your machines apart.

02

Point your tool at the server

Feedmind runs a Model Context Protocol server. Your tool needs its address and your token. The help article walks through it without naming anyone’s product.

03

Ask in words

Your tool decides how to draw the answer. What we control is which numbers it gets and the caveats that travel with them.

What a token can read

Stated in plain words, because scope strings are not an explanation.

Your catalog and product scores with the four dimensions. Your products, their issues and the fixes proposed for them. Your insight cards with their numbers. Your proof metrics: impressions, clicks and Merchant Center status as read from Google.

ShownOnce
StoredAs a hash
ScopeRead-only
RevocationEffective on the next request
AuditCreated and last used, per token
No operation exists that writes
It cannot do anything else: no billing data, no addresses, no sign-in details, no other merchant’s anything. The server exposes nothing writable, so a token cannot approve a fix, push to your store, or change a setting. Revoking one takes effect on its next request.
Security

A token is a credential, and we treat it like one.

It is shown once and never recoverable. Every token lists when it was created and when it was last used, because a token nobody uses is a token to revoke. The server is a public surface with its own security review before it ships, held to the same fail-closed rule as everything else we expose.

Shown
Once
Stored
As a hash
Scope
Read-only
Revocation
Effective on the next request
Audit
Created and last used, per token
Cost
Included on every plan, including Free

AI tool access is not a paid extra. Every plan, including Free, can create a token and read its own data.

What you are responsible for

Anyone holding your token can read your Feedmind data. Treat it like a password and revoke it if you are unsure.

Make your own data as readable as we ask your catalog to be.

Create an account, connect your store, and the token is one screen away in settings.