MCP Connectors, your company's data inside the AI you already use
We connect Claude, ChatGPT or Gemini to Ports & Shipping and Stonify OnManage. People ask in natural language and the answer comes from the company's system, with the permissions of the person asking.
Where is container MSKU 123456-7 and when does it arrive in Luanda?
tool: get_container_status · Ports & Shipping · read-only
It is on board the vessel Atlântico Sul, which left Lisbon on 28/10.
Expected arrival in Luanda on 9/11. B/L LIS-LAD-0412, not yet invoiced.
Which marble blocks do we have in stock over 2 m³?
tool: find_stock · Stonify OnManage · read-only
Three blocks: B-1187 (3.1 m³), B-1203 (2.6 m³) and B-1210 (2.2 m³).
B-1203 is reserved for project O-58.
- 01
In the AI you already use
Claude, ChatGPT or Gemini, on the plans that allow connecting MCP servers.
- 02
Each person's permissions
The connector only shows what the person can already see in the system.
- 03
Read-only by default
You ask and look things up. The connector writes nothing to the ERP.
Who it is for
Operations
Knows where a container is or which port calls are due this week without opening screens or asking anyone.
Sales
Answers the customer with the cargo status or the stock available right now.
Management
Asks about margins, costs and volumes and sees where each figure comes from.
IT team
Decides who uses the connector, with which AI assistant, and sees the log of questions.
The question comes from the AI, the answer comes from the company's system
The AI assistant calls the connector, the connector queries the system with the permissions of the person asking and returns only what that person can see.
- Claude
- ChatGPT
- Gemini
on the plans that allow connecting MCP servers
- Sevways MCP Connector
- permissions of the person asking
- read-only by default
- log of every question
- Ports & Shippingon Sage X3
- Stonify OnManagenatural stone
Two connectors, one per vertical
Each connector knows the system it connects to and answers with the data held there.
Ports & Shipping
For shipping agents and freight forwarders on Sage X3. Examples: "where is the container and when does it arrive?", "which port calls do we have this week?", "which B/Ls have not been invoiced yet?", "what was the margin on this port call?". In production.
Stonify OnManage
For the natural stone industry. Examples: "which marble blocks do we have in stock over 2 m³?", "how much did this block cost up to the slab?", "which slabs are reserved for this project?". In production.
Security and control
The data stays in the company's system
The connector queries the ERP when there is a question and returns only the answer. The database is not copied elsewhere.
The permissions of the person asking
Each person signs in with their own user. The connector sees no more than that person sees in the system.
Nothing is written without a person
The connector only reads. Writing to the ERP goes through an agent, with a person approving each proposal.
Every question logged
Who asked, what the connector looked up and when. The log can be followed in MAJISTI.
From the first connector to daily use
- Choosing the questionsthe questions the team asks every day and the data that answers each one
- Connection to the systeminstalling the connector and connecting it to Ports & Shipping or Stonify OnManage, with the company's users
- Test with a small groupa few people use the connector in the company's AI assistant and we review the answers
- Daily usethe connector is opened to more people and new questions come in later versions
Connected to
The systems where the connector looks up data and the next steps in the Gen AI offer.
Ports & Shipping
Port calls, cargo, documentation and invoicing for shipping agents
View product →Stonify
Natural stone, from quarry block to delivered project
View product →AI agents
Agents connected to the ERP, with human validation
View product →MAJISTI
AI control tower: governance and observability
View product →
See your data in the AI you already use
We show you the connector in a session, with the questions your team asks every day.
Got a project? Let us talk.
Answer within 24 working hours.


