Myriam is typically a Command Line Interface (CLI) tool. Here are the standard workflows:

Myriam includes modules for static analysis without running the app. You can decompile the code to view the logic:

At its core, Project Myriam rejects the prevailing "one-to-many" model of AI, where a single model like ChatGPT or Gemini serves billions of users with generalized knowledge. Instead, it champions a "one-to-one" paradigm. Myriam is an AI that, from its inception, is trained exclusively on the biometric, psychological, and behavioral data of its sole user. It learns not from the entire internet, but from the entire life of its partner: their sleep patterns, stress responses in voice memos, writing style in private emails, heart rate variability during work, and even subconscious eye movements while reading. This narrow, deeply personal training data serves two crucial purposes. First, it creates an AI of unparalleled predictive accuracy regarding the user’s needs and emotional states. Second, it acts as a natural safety constraint: Myriam cannot be weaponized against society or copied to serve another master, because its entire intelligence is a unique reflection of a single, irreplaceable human. In essence, Myriam is as fragile and unique as the person it mirrors.

Project Myriam's core mission revolves around three primary objectives: to provide comprehensive care and support, to empower individuals and families, and to foster community resilience. To achieve these goals, the project is structured around five key pillars: