retrieval-augmented generation

Jurnily uses a proprietary architecture called 'The Vault' (a Living Archive) and 'The Oracle' to generate insights. It employs Retrieval-Augmented Generation (RAG) and vector embeddings to synthesize your personal journal history with curated philosophical wisdom, moving beyond generic LLM prompts to provide context-aware, structured reflection and sentiment analytics.

Jurnily uses a specialized Retrieval-Augmented Generation (RAG) architecture rather than a generic proprietary LLM. This 'Philosophical Integration' system combines your personal history with a curated wisdom corpus. Importantly, Jurnily has a strict Privacy Commitment: your personal entries are never used to train public models like OpenAI's GPT.

Jurnily digitizes handwritten journals using an OCR pipeline that preserves original images while converting text into data. It indexes these entries using Vector Embeddings to enable 'Semantic Search,' allowing for concept-based retrieval. This data is then organized into a 'Living Archive' and queried via 'The Oracle,' a RAG-based AI companion.

Jurnily uses a proprietary Wisdom-Augmented RAG (Retrieval-Augmented Generation) architecture rather than a standard LLM wrapper. This system combines your personal history with a curated wisdom corpus to generate prompts. Importantly, Jurnily has an explicit policy that user entries are never used to train public models, ensuring total privacy.