Let’s organize an ARBoard demo

Use the “Book a demo” button, which opens the calendar on Calendly, or write to info@arboard.eu or directly to Giorgio Insinga, CEO, at giorgio@arboard.eu. In the demo we show you every feature, answer your questions and think together about how ARBoard can fit into your company’s work.

For teams that handle PCB diagnosis, repair and rework: technicians at the bench, experts in other departments and whoever prepares documents for quality or for the customer. ARBoard brings the board, the information and the people together in a single workspace, with an AI colleague at their side.

ARBerto is ARBoard’s AI colleague. It reads test reports and assigns each failure to the right component. You can ask it by voice or in writing: it searches manuals, datasheets and history, cites its sources and shows the answer on the board, highlighting components and nets. It runs locally, on the company’s servers.

The test report enters ARBoard and the AI turns it into data. ARBoard counts how many times each component has made a test fail, and the components on the board are colored accordingly: the fault map shows where to start. Next to the failure is the procedure used last time.

For the board, a single ODB++, IPC-2581 or Gerber file is enough: components, nets, layout and layer stack-up. The 3D model comes from a STEP file and the bill of materials from Excel or CSV; manuals, schematics and datasheets are uploaded as PDF. For the history, the AI imports the reports you already have: it reads PDF, Word, Excel, CSV, XML and free text, as well as AOI, ICT and flying probe exports.

In the same session, on the same board. The technician asks for assistance, stating the urgency and the problem; the expert, even from another department, receives the request and joins the session with “Claim & Join”. They talk in the built-in chat, and when the expert selects a component, it lights up on the technician’s screen too.

The Work Order records every action, with author and time, and when the case closes the report is already written: AI summary, notes and action log, exported to Word, Excel or CSV in one click. The session becomes a repair example, searchable and reproducible step by step, and the component catalog adds up the failures of every board: the next similar case starts from there.

Yes. ARBoard runs entirely on your servers: application, database and AI models. No data leaves the company network and there is no telemetry; in the local configuration ARBoard also works without an Internet connection. How to verify it with your IT team is explained on the ARBoard on-premise page.

Yes. ARBoard connects to models through the OpenAI-compatible API: you can use an open-source LLM you already run or provide a new one, on an engine that supports the features ARBoard uses, for example vLLM, Ollama or llama.cpp. “OpenAI-compatible” only names the API format: the model and the data stay in the company. Engines and requirements are on the on-premise installation page.

A server with Docker for Linux containers, space for projects, documents and history inside your backup, the Wibu-Systems CodeMeter USB license key, workstations with a supported browser, the network (LAN, HTTPS and WSS, a VPN for remote users) and compute and memory suited to the chosen AI model. The application server and the AI resources are sized separately, together with your IT team, with a test on your real load. All the details are on the on-premise installation page.