Djournl

Summary

Djournl is a journaling app that treats a mood log as data worth reading back. Entries are private by default, insights are generated weekly, and the AI provider is swappable through an environment variable.

Problem

Most journaling apps store entries and stop there. The value in a journal is the pattern across weeks, and reading back through dozens of entries to find that pattern is work nobody does.

Approach

A vanilla HTML, CSS, and JavaScript frontend keeps the client small and the behavior explicit. An Express API in front of SQLite holds the data. A single provider abstraction lets the deployment point at Anthropic, OpenAI, or any OpenAI-compatible endpoint by changing one environment variable, so the app is never locked to a vendor.

Outcome

Daily and check-up journaling, a mood dashboard, weekly AI pattern insights, shared streaks with invite codes, tags, favorites, photo notes, JSON backup and restore, a guided feature tour, and account recovery through a personal question.

Honest limitations

Signup is open, so the live deployment is shared rather than invite-only. The weekly insight quality depends entirely on the configured model, and there is no offline mode. The live link is the source of truth for current behavior.

Detail

Security and quality

scrypt password hashing, HttpOnly session cookies, rate limits, sanitized input, CSP headers, and HTTP-level integration tests.

Deployment

Docker, with guides for Fly.io, Railway, and Render.

Data ownership

Full JSON export and restore, so an entry list is never trapped in the app.