Both are AI-native. The architecture is different.
Reflect is a polished networked note app with AI built in. GPT-4. Backlinks. Graph view. End-to-end encryption. It's designed for individual thinkers who value privacy and connected ideas.
Hjarni is simpler on purpose. No bundled AI. A built-in MCP server lets the AI of your choice read, search, and write your notes directly.
AI inside versus AI alongside
Reflect's AI lives inside the app. You ask it questions and it works with your notes. But the AI stays in Reflect. Your notes don't travel to Claude or ChatGPT.
Hjarni takes the opposite approach. Your notes are available to any AI that supports MCP. Today that's ChatGPT and Claude. Tomorrow, whoever adds MCP support next.
A concrete workflow
You're synthesizing research from 20 notes. In Reflect, you'd use the built-in AI to summarize and link related ideas. The graph view helps you see connections visually.
In Hjarni, you'd ask Claude to search your research folder, synthesize across the notes, and save the summary as a new note. You can set folder-level instructions: "Cite specific notes. Use academic tone." Different path. Both get you to synthesis.
Reflect connects your ideas. Hjarni connects your ideas to your AI.
When Reflect is the better fit
If you want end-to-end encryption and AI built into a beautiful personal note app, Reflect is excellent. It's especially strong for solo thinkers who don't need team features.
When people choose Hjarni instead
The case for Hjarni starts when you want your notes available to ChatGPT or Claude directly. When you want team collaboration. When you want different AI instructions for different folders. Or when you already pay for an AI and don't want another subscription.