The follow of journaling is quickly vanishing. At one time, the leather-bound diary or the composition e-book served as a repository for all the things from mundane day by day observations to the deepest existential crises. Recording these ideas and observations required an individual to take a seat in silence, mirror on their experiences, and commit their inner world to paper.
At the moment, the examined life has largely migrated from the non-public web page to the general public social media feed. With this modification, writing has develop into extra performative than reflective. The deliberate tempo of handwriting has additionally been sped up by the quick tempo of digital instruments. Briefly, the follow of self-reflection is fading, and that isn’t good for any of us.
The {hardware} (📷: cristal-orion)
GitHub consumer cristal-orion nonetheless sees journaling as an vital follow, and has developed a distraction-free digital instrument to get probably the most from it. Referred to as Reminor, it’s an AI-powered private diary. As you write your ideas, Reminor analyzes them, remembers them, and chats with you about them. All of that is that can assist you get extra out of your journal entries — issues that could be misplaced within the a whole bunch of pages that accumulate over time.
Not like most fashionable AI journaling platforms, Reminor is just not a web site or a subscription service. It’s designed to run domestically on a Raspberry Pi. The system hosts its personal internet interface and backend, storing each entry straight on the system. Customers kind on an exterior keyboard, and cristal-orion used a 4-inch Waveshare LCD mounted inside a 3D-printed case, making a devoted writing equipment that sits quietly on a desk and consumes beneath 5 watts of energy.
Underneath the hood, the Pi runs a Docker-based software program stack composed of a Svelte frontend and a FastAPI backend. Journal entries are saved as easy dated textual content recordsdata. Reminor processes every entry utilizing native embedding fashions that convert writing into mathematical vectors — numerical representations of which means. These vectors are saved utilizing a system referred to as Memvid, which compresses them into video frames to preserve storage whereas enabling semantic search. As an alternative of trying to find particular phrases, a consumer can ask questions like “when was I fearful about work?” and retrieve associated passages written months or years aside.
An in depth-up of the system (📷: cristal-orion)
The system additionally tracks emotional tone throughout eight dimensions, visualizing long-term temper developments and writing streaks. An built-in chat interface permits the consumer to ask questions on their very own previous; the software program retrieves related entries and makes use of an AI mannequin to summarize patterns or connections. Importantly, all of this could run offline utilizing native fashions. Distant massive language mannequin APIs from suppliers similar to OpenAI, Anthropic, or Google are elective and solely used if the consumer explicitly permits them.
For those who’d wish to check out Reminor for your self, begin by trying over the GitHub repository.—
