Activity
Length by Year
LLM Insight
Projected cost unavailableThe selected model reads diary text when updating missing or stale entry cache records. Final insights use cached capsules and yearly summaries.
Cache status unavailable.
Standing Topics
Last 30 daysPrivate by Design
Entries are decrypted only while this page is unlocked. Locking clears the key and readable entries from memory and unloads the local model. LLM actions run only when you request them. Local inference keeps diary text in this browser. OpenRouter requests send the selected context with Zero Data Retention routing and provider data collection denied.
LLM
Pricing loads from OpenRouter for the saved model.
Qwen3.5 4B runs in this browser with no API key or per-request charge. Save the Local provider, then request guidance, cache updates, or insights. The model loads automatically when needed.
The download is several GB and requires a WebGPU-compatible desktop browser with enough GPU memory. Runtime files and model weights download from a CDN and Hugging Face on demand. Your diary stays in this browser during inference.
Model files are cached by the browser outside your encrypted vault. Cancel or unload to release GPU memory; locking also unloads the model. Downloaded files remain cached for reuse.
Model not loaded.
LLM settings and cached insights are encrypted with this vault. Diary entries may be mixed-language; responses use the configured language.
Prompt Templates 8 editable messages
Available placeholders: {responseLanguage}, {promptText}, {year}, {entryCapsuleFormatInstructions}, {insightReportFormatInstructions}. Keep {promptText} in user messages.
Appearance
Backup
Backups are encrypted. Decrypted Standard Notes JSON can be imported into this unlocked vault. Plain TXT exports are unencrypted.