It demonstrates that a functional, private, offline AI operating shell can run on extremely modest hardware (4 GB RAM, 8 GB USB) using only free and open-source components, challenging the assumption that local AI requires expensive infrastructure.
Bootable Model As System
Boot directly into an AI operating shell.
BMASS is an open-source experiment that treats a local language model as the primary interface to a computer rather than another application running inside one.
Instead of booting a desktop and then opening an AI program, BMASS starts directly into a lightweight Linux environment where the AI becomes the operating interface.
The project explores a simple question:
What if the model was the system?
🤝 Become a Contributor: Post your GitHub username in Discussions and tell us how you'd like to help
23 July 2026
BMASS has undergone its largest architectural change since the project began.
The previous prototype relied on Python orchestration around the language model. That layer has now largely been removed. BMASS is moving towards becoming a native Unix resident compiled directly from C++ using the llama.cpp libraries.
Rather than acting as a conventional chatbot, BMASS now follows a perception–action loop:
User Request
↓
Model
↓
Shell Action
↓
Linux
↓
Observation
↓
Model
↓
Final Answer
The BMASS runtime has been recompiled with a shell execution layer. When the model determines that operating system information is required, it emits a ... action. BMASS executes the command, captures the real output from Linux, feeds that observation back into the model, and allows the model to continue reasoning using actual system state rather than assumptions.
This removes a substantial amount of orchestration code while moving the intelligence closer to the operating system itself. The architecture is intentionally simple and is designed to evolve incrementally.
bmass executablebmass.cpp sourcesystem-prompt.txtThis remains an experimental research project.
The implementation is evolving rapidly and has only been tested on the reference BMASS environment.
Please check file paths before building or running, as your installation layout may differ.
Bug reports, pull requests and architectural suggestions are very welcome.
Thank you to ev
Bootable Model As System (BMASS) – an offline-first AI operating environment
Similar projects matched by category, topics, and programming language.
Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.
A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.
mimic captures traffic from any iOS or web app and automatically generates a Python client library that lets you call the app's API like a regular library.