Vetis

Vetis is best understood as a capture layer, not a journal. The pitch on the site is blunt: "Keeping a journal shouldn't feel like homework." Every design choice follows from that. You get one input for all three capture modes — text, voice, photo — and no categories or templates to pick. You drop a fragment, AI follows up only when a question adds future value, and it ends with "✓ Saved" rather than an open chat thread. That last detail matters: this is explicitly not a ChatGPT-style companion, and the site says so directly ("ChatGPT keeps the conversation going. Vetis helps you finish a record worth saving."). What it actually produces from a single capture is broader than most journaling apps: a worked example on the homepage shows one input ("Today I had lunch with David. Spent $42. He suggested hiring Emily. I felt much less stressed. Need to book dentist.") fanning out into a journal entry, a $42 Food expense, a "Book dentist" task, linked People (David, Emily), a Health mood note, and a Recruitment project thread. That fan-out is the real product. Reviews, dashboards, and natural-language recall all sit on top of it. The search demos — "When did I last meet Alex?" returning March 18 Coffee with discussion context, "How much did I spend eating out this month?" returning $286, "How has my mood changed recently?" returning stress down and sleep improving — are answered from your own captures, not guessed. Strengths: the friction argument is real and the capture UX backs it up; local-first architecture with your device as the source of truth is a genuine differentiator versus cloud-journal apps; de-identification (names replaced with tokens before the LLM call, raw text not stored as a content database) is a concrete, checkable privacy claim rather than a vague promise; custom AI models cost zero credits, so power users aren't penalized for tuning. Weaknesses: Android-only, distributed as a direct APK download, so iPhone, iPad, and Mac users are out entirely and there's no web or desktop access described anywhere on the site. That APK distribution also means no Play Store update path or store-level vetting. The underlying AI model is never named, so you can't judge which foundation model is doing the structuring. The credit system is the gate: the free tier's 200 monthly credits cover roughly 100 capture conversations, and every structuring call, follow-up, and search presumably draws on that same pool, so heavy use pushes you onto the $10 one-time 1,000-credit pack, the $9.9/mo plan (1,500 credits plus the 200 free), or the $99/yr plan (18,000 credits). Teams and couples are out — there's no shared archive. And if you like folders, tags, and templates, the no-structure capture model will feel like a loss of control rather than a relief. Where it fits: solo Android users with high capture volume and low patience for filing — builders logging ideas and decisions mid-flow, professionals tracking meetings and follow-ups across months,