Learn.xyz

Learn.xyz — Generative-AI Micro-Learning & Course Builder (Consumer “Curiosities” + Enterprise Doc-to-Lesson LMS, Active, Not Shut Down)

Not shut down.​ Learn.xyz (learn.xyz, San Francisco) founded 2021 by Christian Byza (CEO) + Arndt Voges (CTO), $3M seed Sep 2023 (Blockchange lead, Coinbase Ventures, Ryan Roslansky, Andrew Barnes), ~19 employees, iOS/Android/web active Aug 2026, no acquire/rebrand/shutdown. Dual-surface product: consumer side = “Curiosities” (ask any question → AI Lumi builds slides+quiz in seconds, Duolingo×Wikipedia vibe, 150+ countries, 60k+ beta curiosities)​ and enterprise side = upload handbook/PDF/deck → 21-sec gamified micro-lesson + LMS with completions/compliance/SCORM/LTI. Category ai-productivity-data; pipeline_stage none. Killer differentiator = turns existing dead docs into 89%-completion micro-lessons with streaks/leaderboards, not another authoring tool you have to write from scratch; not Khanmigo (Socratic tutor, no doc→course), not Mini Course Generator (creator-monetize-first, thinner LMS/compliance), not Coursebox (AI LMS but weaker consumer social loop), not Duolingo (fixed curricula, no custom doc ingestion). 2026 trap: free tier caps 5 seats / 10 lessons (team-only framing, consumer app is separately free with soft limits), self-serve $15/seat/mo tops out at 100 users (bigger orgs forced to Turn-key sales), AI factual drift in consumer “Curiosities” needs human review before enterprise publish, no public API/MCP, gamification can mask shallow retention, founder-profile sites conflate it with a language app (it’s topic-agnostic).

1. One-line positioning

Learn.xyz = dead docs → addictive micro-courses: consumer opens iOS/Android → type “how does QE work” → Lumi generates 5-slide Curiosity + quiz + streak in ~10s, multi-language, social feed shows friends’ topics → enterprise admin uploads 20-page PDF/handbook/deck → AI restructures into 6-8 micro-lessons with checks between chunks → edit in browser before publish → learners get mobile/desktop web app, gamified quizzes, monthly leaderboard with prizes → admin dashboard shows 89% completion (vendor-reported, Mar 2025 58 lessons/user/mo), compliance export, SCORM/LTI on Turn-key, US/EU data region, SOC 2 Type II, GDPR. Fits ai-productivity-data; pipeline_stage none (source→restructure→gamify→assign→complete→report). For HR/L&D rolling out policy/onboarding without rebuilding Slide decks, and for casual learners wanting Wikipedia-depth in Duolingo-format, it clicks; for university-credit courses, deep adaptive tutoring, or creator-monetize storefronts, it’s the wrong shape.

2. 核心功能 Core features

3. Pricing (2026-08, USD)

pricing_type = freemium-per-seat-plus-custom-enterprise; starting_price = $0; free_tier = "yes; consumer app fully free with soft limits; team free = 5 seats / 10 lessons / mobile+web / gamified quizzes, no SSO-SCORM-branding"; pricing_note = "learn-xyz-sf-2021-byza-voges-3m-seed-2023-blockchange-coinbase-roslansky-active-2026-no-shutdown; consumer-free-separate-from-team-free-5-seat-10-lesson-cap; self-serve-15-seat-100-user-cap-then-turnkey-sales; pro-sub-not-include-api-no-public-api-mcp; vendor-claims-21s-per-20pg-89pct-completion-unaudited"

4. Access Type

access_type: mobile-app-plus-web-no-desktop-no-api

access_display: 📱 iOS + Android (consumer Curiosities + enterprise learner) · 🌐 learn.xyz web (builder + learner + admin) · no desktop Electron, no public REST API, no MCP, no browser extension

5. Reviews (2026)

Praise: doc→micro-lesson in seconds kills Slide-rebuild tax, gamification (streak/leaderboard) actually moves completion vs passive PDFs, dual consumer+enterprise surface unusual, free 5-seat pilot is low-friction, multi-lang Curiosities good for global casual learning, SOC2+GDPR+SCORM covers procurement basics.

Gripes: 89% completion / 21s / 58 lessons-per-user are vendor metrics, not independently audited, consumer Curiosities hallucinate on niche topics (needs review gate), free team tier 5 seats/10 lessons too small for real pilot, self-serve 100-user cap forces sales call mid-growth, no API/MCP blocks embedding in existing LMS, gamification can reward quiz-guessing not comprehension, founder-profile sites mislabel it “language app” (it’s topic-agnostic, not Duolingo-clone), LMS analytics shallower than Docebo/TalentLMS.

6. Best for / Not for

7. Competitors