Skip to content

Adaptive Learner

Learn the way you actually learn.

Adaptive Learner is an open-source learning companion built on a research-backed six-method model. You take a 12-question assessment, the app discovers which methods suit you, then AI-supported sessions walk you through a seven-step learning cycle. The app adapts how it teaches based on how you actually learn.

Try it now GitHub

What makes it different

Six methods, not one

Most learning apps pick one approach — flashcards, video, gamified streaks — and assume everyone learns the same way. Adaptive Learner ships six methods (deductive, inductive, error-based, dialogic, contextual, AI-adaptive) and helps you switch between them as you grow.

The six methods →

A seven-step learning cycle

Each session walks through Input → Attempt → Error → Feedback → Adapt → Repeat → Integrate. A dual-prompt AI evaluates per turn whether you're ready to advance, stay, or step back. No conveyor belt — real cognitive pacing.

The seven-step cycle →

Local-first, AI-powered

Toggle between Local mode (everything in your browser, AI calls direct to Anthropic / OpenAI / Gemini) and Server mode (FastAPI backend). Bring your own AI key. Install as a PWA — works offline for past sessions and Dashboard.

Getting started →

Git for learning

Sessions are commits. Trends emerge across weeks. Streaks matter but aren't the point. The point is patterns: which method works for which topic, where you spend cognitive time, what method-switch unlocked progress.

Tracking →

Quick start

  1. Open the live app at astrapi69.github.io/adaptive-learner.
  2. Pick your language + onboard your learning project (topic, goal, timeframe).
  3. Take the 12-question assessment (~2 minutes).
  4. Add your AI API key (Anthropic, OpenAI, or Gemini — free tiers work).
  5. Start your first session from the Dashboard.

Full getting-started guide →

Documentation

Status

Active development. v0.8.0 was released 2026-05-19.

  • 1312 tests (447 backend + 478 plugins + 387 frontend Vitest + 8 Playwright smoke)
  • 8 languages (DE / EN / ES / FR / EL native; PT / TR / JA scaffolded)
  • 7 plugins (assessment / 3 AI providers / session / tracking / tools)
  • 2 storage modes (Local IndexedDB / FastAPI backend)
  • MIT licensed

Source code, issues, and contributions: github.com/astrapi69/adaptive-learner.