Launching today
Writed
Chat with your journal. Nothing leaves your browser
4 followers
Chat with your journal. Nothing leaves your browser
4 followers
Most journal apps say "encrypted." Writed keeps your journal entirely in your browser, with AI that runs on your device. No account, no cloud, no tracking. Ask, "What did I decide about the job offer?" and get answers with citations from your own entries. Export anytime as Markdown. Free and open source. Pro syncs to your own Google Drive. We run zero servers.






ReadHere
Hey Product Hunt π
I wanted a journal I could actually ask questions, without sending my private thoughts to the cloud. So I built Writed.
Your journal stays in your browser, and the AI runs on your device. No account, no upload, no tracking. You can even open DevTools and see that nothing leaves your device.
Everything is stored as plain Markdown, works offline, and you can export anytime. Optional Pro adds Google Drive sync and cloud AI using your own API key.
I'd love to hear your thoughts: would you trust an on-device AI journal? π
on-device is the right call for trust, but it flips the failure mode - instead of worrying about a company leaking my data, I'd be worrying about clearing browser storage or switching machines and just losing years of entries with no cloud copy anywhere. is export-as-markdown meant to be the backup discipline users are expected to build themselves, or is there some local-first sync (not cloud AI, just storage) planned for people who don't want to remember to export?
ReadHere
@galdayanΒ Thanks for your thoughtful comment! I completely agree. That's why the Pro version will include an option to securely sync your journal with your own Google Drive, so you can access it across devices while keeping full control of your data.
Congratulations on the launch! qq How does the local model hold up when the journal grows large and you're asking it to recall across months of entries?
ReadHere
@alieksiaΒ Thanks! The plan is to support both locally hosted models (via llama.cpp or Ollama) and cloud models using your own OpenAI or Anthropic API key. For larger journals, the Pro version will add RAG-based retrieval, so only the most relevant entries are sent to the model instead of the entire journal.