Artie: An Offline AI Archaeology Assistant Built with Raspberry Pi and PiSugar

What if archaeologists could carry an AI assistant in their pocket—even at excavation sites with little or no internet connection?

That is the idea behind Artie, a pocket-sized offline AI archaeology assistant developed by Arteque, a technology and creative education team based in Paris, France.

Built on PiSugar's open-source Whisplay AI Chatbot project, Artie transforms a Raspberry Pi-based AI device into a practical field tool for archaeology. Archaeologists can photograph an object as soon as it is discovered and describe the discovery using their voice. A Raspberry Pi AI HAT+ 2 provides local AI acceleration for analyzing the captured images, while the offline AI workflow organizes and stores the collected information and helps turn it into a structured archaeological report.

The project was created for the 2025–2026 FIRST LEGO League UNEARTHED season and was presented at the 2026 FIRST LEGO League Asia Open Championship in Beijing, held from August 13 to August 16, 2026.

Representing France, the Arteque team received both the Engineering Excellence Award and the Peer Award.

Beyond the competition, Artie demonstrates something even more interesting: what becomes possible when Raspberry Pi, edge AI, computer vision, voice interaction, and open-source hardware are combined to solve a real-world problem.

The Challenge: Bringing AI Where the Internet Doesn't Reach

Modern AI can understand images, transcribe speech, organize information, and generate reports.

There is just one problem: many AI services depend on the cloud.

For archaeology, that can be a serious limitation.

Excavations do not always happen somewhere with fast Wi-Fi or reliable cellular coverage. Archaeologists may be working in remote fields, mountains, ruins, caves, or other locations where connecting to an online AI service is difficult or impossible.

At the same time, field documentation is an essential part of archaeological work. When an object is discovered, researchers need to record not only the object itself, but also information about the circumstances and context in which it was found.

This creates an interesting engineering question:

Could a small AI device help archaeologists capture, understand, and organize field information without depending on an internet connection?

Arteque's answer was Artie.

Meet Artie: A Pocket-Sized Offline AI Assistant for Archaeologists

Artie is designed around a simple field workflow.

When archaeologists discover an object, they can use the device to take a photograph and then describe the discovery by voice.

Instead of immediately stopping to open a laptop, launch software, upload photographs, and type notes, they can record the information directly at the point of discovery.

Artie can then process the collected information locally.

Its vision system analyzes the captured image, while the researcher's spoken observations provide additional context. The AI workflow brings these different sources of information together, organizes them, and saves them locally.

As more observations are collected, Artie can help transform the raw field data into structured records and ultimately generate an archaeological report.

In simplified form, the workflow looks like this:

Discover → Photograph → Local Image Recognition → Voice Notes → AI Organization → Local Storage → Generate Report

The goal is not to replace the expertise of archaeologists. Artie acts as a small digital field assistant, reducing the friction between discovering an object and properly documenting it.

And because its core AI capabilities can operate locally, the workflow can continue even when the internet disappears.

Giving Artie Local Vision with Raspberry Pi AI HAT+ 2

Photographs are an important part of archaeological field documentation, so Artie needs more than voice interaction and a language model.

It also needs to see.

To provide local AI acceleration for image analysis, Artie uses the Raspberry Pi AI HAT+ 2.

When an archaeologist photographs an object, the AI accelerator can help process and analyze the image directly on the device. The resulting visual information can then be combined with the archaeologist's spoken description and other field records.

This means Artie can work with multiple types of information at the same time:

what the archaeologist sees, what the camera captures, and what the archaeologist says.

A field record can therefore include information such as:

Object Photograph + AI Image Analysis + Voice Description + Field Notes

The AI can then associate and organize these different pieces of information before storing them and using them to create a more structured report.

This is particularly important for offline applications.

Instead of sending every archaeological photograph to a cloud AI service for analysis, local AI hardware makes it possible to perform computer vision tasks directly at the edge.

For remote fieldwork, this is one of the most practical advantages of edge AI.

Built on the PiSugar Whisplay AI Chatbot

Artie was developed from PiSugar's open-source Whisplay AI Chatbot project.

Whisplay AI Chatbot began as an experiment in building a tiny voice-enabled AI device around Raspberry Pi. The basic interaction is intentionally simple: speak to the device and receive an AI response through its display and speaker.

The platform combines Raspberry Pi computing with Whisplay HAT and a PiSugar battery system to create a compact, standalone AI device.

Whisplay HAT integrates several of the components needed for a portable AI interface, including:

• a compact LCD display;
• microphone input;
• onboard speaker;
• physical button interaction;
• RGB status indication.

The open-source Whisplay AI Chatbot software expands this hardware into a flexible AI platform with support for speech recognition, large language models, text-to-speech, image understanding, retrieval-augmented generation (RAG), and different local or cloud AI backends.

Arteque took this general-purpose platform and adapted it to a very specific application: archaeology.

Raspberry Pi + AI HAT+ 2 + Whisplay + PiSugar

One of the most interesting parts of Artie is how several different technologies work together, with each component serving a different role.

The Raspberry Pi acts as the main computer, running the software, managing data, and coordinating the overall AI workflow.

The Raspberry Pi AI HAT+ 2 provides dedicated local AI acceleration, helping Artie analyze images captured at the archaeological site.

The Whisplay HAT provides the human interface—including the display, microphone, speaker, buttons, and status indicators—allowing researchers to interact naturally with the device.

The PiSugar battery system provides portable power, allowing the complete system to operate away from a desk or fixed power source.

Together, these components create something more than a Raspberry Pi chatbot.

They create a portable multimodal offline AI assistant designed around a real field workflow.

In simple terms:

Vision lets Artie see. Voice lets researchers record. AI helps understand and organize. Local computing allows it all to work offline.

Why Offline AI Matters in the Field

Artie highlights a broader trend in AI development: moving intelligence from the cloud to edge devices.

Cloud AI models are extremely powerful, but local AI offers several important advantages for field applications.

No Internet Required

The most obvious advantage is connectivity.

If AI models and processing tools can run locally, the device can continue operating when Wi-Fi and cellular networks are unavailable.

For archaeologists working at remote excavation sites, this can make AI tools useful in places where conventional cloud assistants cannot operate reliably.

Local Image Recognition

Computer vision traditionally often involves sending an image to a remote server for processing.

With dedicated edge AI hardware such as the Raspberry Pi AI HAT+ 2, image analysis can instead happen directly on the device.

This makes visual AI much more practical in environments with limited connectivity.

Data Can Stay on the Device

Local processing also reduces the need to continuously upload photographs, recordings, and field notes to external servers.

For workflows involving unpublished discoveries or potentially sensitive research information, having the option to process and store data locally can be valuable.

A More Natural Field Workflow

Dedicated hardware also changes how people interact with AI.

Instead of:

Laptop → Software → Upload → Type → Process

the interaction can become:

See something → Take a photo → Speak → Continue working

That difference may look small, but in a field environment it can completely change how useful an AI tool feels.

From a Chatbot to a Purpose-Built AI Tool

One of the most interesting aspects of Artie is that the underlying platform was not originally designed specifically for archaeology.

Whisplay AI Chatbot is an open platform.

The same combination of a Raspberry Pi, camera, microphone, speaker, display, battery, and local AI can potentially become many different devices depending on the software and AI models running on it.

Arteque turned it into an archaeology assistant.

Another developer could use a similar platform to create a field research assistant, maintenance tool, offline translator, educational device, wildlife observation recorder, or portable knowledge assistant.

This is one of the reasons edge AI on Raspberry Pi is so interesting.

Instead of AI existing only inside a browser window, developers can give it a physical interface designed around a specific task.

Artie demonstrates that transition perfectly:

from a general-purpose AI chatbot to a specialized AI tool.

Built for the FIRST LEGO League UNEARTHED Season

The timing of the project was especially appropriate.

The 2025–2026 FIRST LEGO League season is called UNEARTHED, an archaeology-inspired challenge encouraging young innovators to investigate problems faced by archaeologists and develop new technological solutions.

The season explores how STEM, engineering, and technology can help us discover, understand, and preserve the stories hidden beneath our feet.

For Arteque, this became an opportunity to investigate a practical part of archaeology:

How can technology make it easier to capture and organize information directly in the field?

Arteque is an educational organization based in Paris whose activities bring together Art, Technology, and Music. Its programs include computer science, engineering, programming, digital arts, music, and hands-on creative projects.

Artie fits naturally into that philosophy.

The project combines product design, electronics, Raspberry Pi development, edge AI, computer vision, voice interaction, and a real-world research problem in a single device.

From Paris to Beijing

Artie eventually traveled far beyond the classroom.

From August 13 to August 16, 2026, the Arteque team participated in the 2026 FIRST LEGO League Asia Open Championship in Beijing, China, representing France alongside teams from around the world.

The event brought together young innovators for several days of robotics, engineering, problem-solving, and international exchange.

For Artie, the UNEARTHED theme was not simply a background story. The challenge asked participants to think about real problems encountered when studying and preserving the past and explore how technology could help solve them.

Artie was Arteque's response to that challenge.

And the team's work was recognized with two awards:

Engineering Excellence Award — recognizing the team's strong engineering work and overall performance.

Peer Award — a particularly meaningful recognition reflecting the impression the team made on fellow participants.

For a project that began with a question about bringing offline AI into archaeological fieldwork, returning from Beijing with two awards was an impressive milestone.

A Small Device with a Much Bigger Idea

For us at PiSugar, Artie is exciting for another reason.

When we created the Whisplay AI Chatbot project, the goal was never simply to build another device for talking to an LLM.

A Raspberry Pi, a tiny display, a microphone, a speaker, a battery, an AI accelerator, and an AI model are all individual components.

The interesting part begins when someone combines them around a real problem.

Arteque did exactly that.

They took an open-source Raspberry Pi AI platform and asked a different question:

What would this become if it were designed for an archaeologist instead of a chatbot user?

The result is a device that can be carried into the field, capture photographs and voice notes at the moment of discovery, analyze visual information locally, organize the collected data, and help transform raw observations into structured documentation.

It is a great demonstration of how offline AI, edge computing, computer vision, multimodal AI, open-source software, and Raspberry Pi hardware can move artificial intelligence beyond the desktop.

Build Your Own Raspberry Pi AI Device

Artie is based on the open-source PiSugar Whisplay AI Chatbot project, giving developers, educators, students, and makers a foundation for exploring their own AI applications.

Depending on the hardware and software configuration, developers can experiment with:

• offline speech recognition;
• local large language models;
• computer vision;
• AI-accelerated image recognition;
• vision-language models;
• local text-to-speech;
• RAG and local knowledge bases;
• voice interaction;
• multimodal AI workflows;
• custom AI tools and agents.
• The result could become an archaeology assistant like Artie—or something completely different.

That is ultimately what makes projects like this important.

The future of AI is not only about making models larger. It is also about finding new places to put intelligence and new problems for it to solve.

Artie shows what can happen when that intelligence becomes small enough to carry in your hand—and powerful enough to keep working even when the cloud is out of reach.

Explore the Project

Whisplay AI Chatbot
PiSugar's open-source Raspberry Pi AI assistant platform for building compact voice, vision, and AI projects.

Whisplay HAT
A compact Raspberry Pi HAT integrating a display, microphone, speaker, buttons, and status lighting for portable AI applications.

Raspberry Pi AI HAT+ 2
Dedicated AI acceleration hardware used by Artie to enable local image recognition and analysis.

Arteque
A Paris-based educational organization combining art, technology, music, programming, engineering, and creative learning.

Artie is an independent project developed by Arteque using PiSugar's open-source Whisplay AI Chatbot platform.