VocabGen — Vocabulary Web App & Flashcard Tool

Look up words, batch-process CSV lists, and study with flashcards — all in your browser. Powered by LLM providers. Free, open-source, runs locally.

Download Latest Release View on GitHub

What is VocabGen?

VocabGen is a free, open-source vocabulary web app for language learners at B2–C1 level. Start the app, open your browser, and you can look up words, upload CSV word lists for batch processing, study with flashcards, and manage your vocabulary database. It uses LLM providers (OpenAI, Anthropic, AWS Bedrock, Ollama) to generate structured vocabulary entries with definitions, translations, connotation, and register. Everything runs locally on your machine — your data stays private.

Features

Web-based vocabulary app — lookup, batch, flashcards, database, and config in your browser
Flashcard study mode — flip cards, rate difficulty, filter by language and tags
Sentence analysis — grammar checking, corrections, vocabulary extraction, and translation
LLM-powered vocabulary lookup — definitions, translations, connotation, register
CSV batch processing — upload word lists via the web UI or process from the command line
Multi-language support — 11 built-in languages, any language via full name
SQLite cache — never pay twice for the same lookup
Single binary — zero runtime dependencies, cross-platform
Docker support — multi-arch images on GitHub Container Registry
Dry-run mode — preview operations before spending API credits
CLI for power users — script lookups and batch jobs from the terminal

Supported Languages

Dutch German French Spanish Italian Portuguese Russian Polish Turkish Hungarian English + any language

LLM Providers

OpenAI

Anthropic

AWS Bedrock

Google Vertex AI

Ollama (free)

Azure OpenAI

LM Studio

Free to try, better with a paid model. The cheapest way to get started is Ollama — install it, pull a model, and you're running locally for free. For the best translation quality — especially for less common languages, connotation/register distinctions, and contrastive notes — use a large model like Claude Sonnet/Opus or GPT-4o. Local models work but produce noticeably lower quality on nuanced vocabulary tasks. Use --dry-run to preview results before committing to a provider.

Quick Start

# Start the web app (open http://localhost:8080 in your browser)
vocabgen serve --port 8080

# Or use the CLI for scripting
vocabgen lookup "maison" -l French
vocabgen batch --input-file words.csv --mode words -l German

# Free local LLM (no API key needed)
vocabgen lookup "casa" -l Italian --provider openai \
  --base-url http://localhost:11434/v1

Use Cases

Language learners preparing B2–C1 vocabulary lists from textbook chapters
Students creating flashcard decks from CSV word lists
Teachers generating structured vocabulary materials for classes
Developers looking for a simple Go web app example (CLI + HTMX + SQLite)
Anyone who wants LLM-powered vocabulary lookup with local caching

Documentation

User Guide — installation, configuration, daily usage
Architecture — system design, package layout, data flows
Deployment — cross-compilation, CI/CD, Docker