Price a range
"Mark up this supplier list 35 percent and load it." It matches the catalogue and prices the lot.
Bring your own agent
Run your shop from the AI you already use.
Connect Claude, ChatGPT, Cursor or any MCP client to your Requex workspace. Ask it to price a range, fill the diary, order stock or pull the month's numbers. The work happens in Requex and it reports back.
Claude
ChatGPT
Gemini
Cursor
VS Code
+ any MCP client
Connect in three steps.
No integration project. No middleware. Just point your agent at your workspace.
Launch a workspace
Sign up and your shop gets its own secure Requex workspace.
Add the MCP server
Paste your workspace's MCP address into your agent and approve the connection.
Ask for the work
In plain language. The agent uses your shop's tools, within your permissions.
What it can do
A few things to ask it.
Fill the diary
"What is quiet this week, and who can I call to fill it?" It reads the diary and suggests.
Order stock
"We are low on the popular 17s, raise the orders." It drafts the POs within your limits.
Pull the numbers
"How did mobile do last month versus the bays?" It runs the report and explains it.
Your agent, your shop, your rules.
The connection runs on the open Model Context Protocol, so you are never locked to one vendor. Requex spends no AI on your behalf: your agent does the thinking, on your account.
- Permission aware. The agent can only do what the connected user is allowed to do.
- Open protocol. Any MCP client connects. Switch agents whenever you like.
- Your data stays yours. Isolated to your workspace, never shared across shops.