Skip to content
Back to home

MCP servers

Let any AI assistant work with your systems.

We build MCP servers that connect Claude, ChatGPT, Copilot or your own agents to the systems you run. Each person gets answers from live data, with their own rights, and every change follows the approval rules you set.

See it on a system like yours

Pick a system
01 · Ask
Which orders from yesterday haven’t shipped yet?
02 · Tool call
orders.search(status: "unshipped", placed: "yesterday")
Signed in as youRead-only
03 · Checks
Card data never leaves the shopLogged
04 · Answer
14 orders. 9 are waiting for stock and 5 for a carrier pickup. Shall I draft a delay email to the 9 customers?
Delay emailWaits for your approval
01 · Ask
Which supplier payments are due this week, and are any blocked?
02 · Tool call
payments.list(due: "this_week")
Signed in as youRead-only
03 · Checks
Account numbers maskedLogged
04 · Answer
23 payments, €182,400 in total. 2 are blocked: a missing VAT number and a bank detail change waiting for review.
Release blocked paymentsNeeds 2 approvers
01 · Ask
Why was line 3 down last night?
02 · Tool call
downtime.get(line: 3, period: "last_night")
Signed in as youRead-only
03 · Checks
No writes to machinesLogged
04 · Answer
47 minutes in 2 stops. The longest, 31 minutes, was a jammed feeder at 02:14. Maintenance ticket MT-882 is open.
Nothing to changeRead-only answer
01 · Ask
Which shipments to Germany are late today?
02 · Tool call
shipments.search(country: "DE", status: "late")
Signed in as youRead-only
03 · Checks
Driver location only as neededLogged
04 · Answer
6 shipments, all on one route delayed by a road closure. The carrier’s new ETA is tomorrow at 10:00.
Customer updateWaits for your approval
01 · Ask
Show my team’s unused licences.
02 · Tool call
licences.list(team: "mine", status: "unused")
Signed in as youCustomer’s own account
03 · Checks
Per-customer limitsLogged
04 · Answer
7 of 40 licences are unused. 4 belong to people who left last month. Shall I release them?
Release 4 licencesWaits for confirmation
An example conversation: a person asks an AI assistant a question, the assistant calls a named tool on the MCP server while signed in as that person, the server applies its checks, and the assistant answers. Changes beyond the allow-listed actions wait for approval.

Why MCP

Build once. Every assistant can use it.

Your API is written for developers. An MCP server turns it into tools any AI assistant can find and use, with sign-in, limits and logging built in.

  • Claude
  • ChatGPT
  • Gemini
  • Microsoft Copilot
  • Cursor
  • Your own agent

An open standard, introduced by Anthropic and governed by the Linux Foundation since December 2025.

Compare

Without MCP

  • Claude
  • ChatGPT
  • Your agent
  • Shop
  • ERP
  • Ledger

Every agent needs its own integration with every system: 9 to build, secure and maintain.

With MCP

  • Claude
  • ChatGPT
  • Your agent

MCP servers

  • Shop
  • ERP
  • Ledger

One server per system, and any agent can use it: one place for sign-in, limits and the log.

Security and control

Agents get rights, not keys.

A demo server is easy. One that touches orders, money or a production line needs the controls your security review will ask for. Ours ship with them.

  • Acts on behalf of a named person

    The agent has its own identity and works with that person’s rights, so every log line shows both: who asked and which agent acted.

  • Read-only by default

    Every server starts read-only. Change tools are added one by one, each with its own limits.

  • A person approves changes

    Beyond the low-risk actions you have allow-listed, every change waits for 1 approver, or 2 where you need four eyes.

  • Masked by policy

    Your policy lists the fields to mask, and the server applies it, not the model. Card data is read only as tokens, never as full numbers.

  • Prompt-injection safeguards

    No model is fully immune, so we limit the damage: the agent only reaches the tools and data you allowed, and changes beyond your allow-list wait for a person.

  • Every call on record

    Who asked, which agent acted, who approved and what came back, in your log and SIEM, masked the same way as answers.

  • Runs in your infrastructure

    The server sits in your environment, next to your systems. Your data does not pass through UAPP.

  • Kill switch

    One step revokes an agent’s mandate or disconnects the server. Running tasks stop and no new calls go through.

How we work

One system first, then the rest.

We start with one system and the questions people ask most, prove the answers are right, and only then add change tools and more systems.

Book a technical call
  1. Step 1Discovery

    Questions and data map

    We list the questions and tasks, the systems behind them and who may see what.

    • Question list
    • Data and access map
    • Tool plan with owners
  2. Step 2Read-only

    First server

    One system, read-only tools, tested against your questions in your environment.

    • Semantic data model
    • Read tools
    • Test results
  3. Step 32-4 months to a pilot

    Changes and more systems

    Change tools with confirmation, then more systems under the same rules and the same log.

    • Change tools with approval
    • One audit log
    • Handover and documentation

Contact

Tell us which systems your agents should work with.

A 30-minute technical call with an engineer. Bring one system and the questions people ask about it; we tell you plainly what an MCP server for it would take.

  • We look at one system end to end
  • You get a data map and a tool plan
  • Confidential from the first call, NDA if you need it

Prefer email? office@uapp.group

We use your details only to reply to you, as set out in the UAPP Privacy Policy.