aethon-oss

AETHON OSS is an open-source reference implementation of a local AI assistant workspace. The core idea is that the AI keeps its memory as plain Markdown files on your own machine rather than storing anything in the cloud, and it uses AI models that run locally through software compatible with Ollama (a tool for running AI models on personal hardware) rather than sending requests to external services.

The project positions itself against the typical AI assistant demo, which tends to be stateless (no memory between sessions) and dependent on cloud APIs. AETHON instead keeps daily logs that link chronologically, similar to how tools like Obsidian connect personal notes. The memory accumulates as readable Markdown so the user can inspect what the assistant knows and how it is organized. The repository includes a small Python package for indexing that memory and doing basic local retrieval, a script for linking daily note files in chronological order, a standalone web page for visualizing how memory entries connect to each other, and example memory files with personal details removed.

The maintainer's actual working AETHON setup contains private notes, browser profiles, API keys, and social account automation that are intentionally excluded from this public version. What is shared here is the structural and tooling layer that could be reused by others who want to build a similar local-first assistant workflow.

The stated goal is to evolve this public extraction into a reusable framework for maintainers, researchers, and tool builders who want durable, inspectable agent memory without relying on cloud services. The project is described as an early extraction, meaning it reflects a working private system but has not yet been polished into a finished, general-purpose tool.

The license is MIT, and setup involves cloning the repository, creating a Python virtual environment, installing the package, and running a command-line scan against the included example memory files.