تناغم About Tanaghum
A free, open-source tool for generating professional Arabic listening comprehension lessons
What is Tanaghum?
Tanaghum (تناغم, meaning "harmony" in Arabic) is a free, client-side Arabic listening comprehension lesson generator. It transforms YouTube videos, audio files, or text into complete, pedagogically-sound lessons with synchronized transcripts, vocabulary lists, and comprehension questions.
Everything runs directly in your browser - no server costs, no sign-up required, no data leaves your device. Generate professional-quality lessons in minutes and export them as standalone HTML files that work offline.
Key Features
- AI-Powered Transcription: Uses Whisper (via Transformers.js) for accurate Arabic speech-to-text directly in your browser
- ILR-Calibrated Content: Lessons are tagged with ILR proficiency levels for appropriate difficulty matching
- Pre/While/Post Framework: Research-backed listening comprehension structure with questions at each stage
- Smart Search: AI-evaluated YouTube search to find suitable content for language learning
- Vocabulary Extraction: Automatic word lists with roots, translations, and definitions
- Standalone Export: Download complete lessons as single HTML files that work offline
- Community Gallery: Share and discover lessons created by other users
ILR Proficiency Levels
The Interagency Language Roundtable (ILR) scale is the standard grading scale for language proficiency in the U.S. federal government. Tanaghum uses this scale to help learners find content appropriate for their level.
| Level | Name | Description |
|---|---|---|
| 1 | Elementary | Can understand short, simple texts with familiar vocabulary. Basic greetings and routine phrases. |
| 1+ | Elementary+ | Can understand simple conversations on routine topics. Recognizes basic sentence patterns. |
| 2 | Limited Working | Can understand factual news, personal narratives, and simple arguments. Follows straightforward discussions. |
| 2+ | Limited Working+ | Can understand standard news broadcasts, formal speeches, and extended arguments on familiar topics. |
| 3 | General Professional | Can understand political analysis, technical discussions, and complex arguments. Follows most broadcasts. |
| 3+ | General Professional+ | Can understand complex argumentation, nuanced content, and subtle cultural references. |
Technology Stack
Tanaghum is built entirely on modern web technologies and runs 100% client-side:
Transformers.js
Runs Whisper speech recognition directly in the browser using WebAssembly
Google Gemini
Google AI Studio (Gemini) for question generation, vocabulary extraction, and translation
Cloudflare Workers
Lightweight proxy for YouTube data and LLM APIs, no server hosting needed
IndexedDB
Local storage for caching transcriptions and lessons without server dependencies
Web Audio API
Client-side audio processing for playback control and visualization
GitHub Pages
Free hosting with zero server costs - just static files
The Pre/While/Post Framework
Tanaghum generates lessons using the research-backed Pre/While/Post listening comprehension framework used by language educators worldwide:
- Pre-Listening: Activate prior knowledge and set expectations. Questions help learners think about the topic before listening.
- While-Listening: Focus on specific information during playback. Questions guide active listening for key details, main ideas, and supporting points.
- Post-Listening: Reflect and synthesize understanding. Questions encourage critical thinking, personal response, and connection to broader themes.
Created by SSgt Jesse Morgan, USMC
Built for Arabic language learners and educators everywhere.
Free and open source under the MIT License.