Langdogu

Learn a language from what you already read, watch and hear.

Langdogu is a language-learning system built around a personal ontology: a map of what you know, what you don't, and how words connect. Instead of a separate app with its own exercises, it turns your everyday reading, video and audio into study material and keeps one record of your knowledge across all of it, built up over years.

Lingua Layer

Lingua Layer is a general-purpose translation extension for Firefox. It translates any web page, the whole page or just the sentences you select, including text inside images and manga, overlaid right where the original text sits. You can view the original, the translation, or both. It currently supports Chinese and Japanese into Korean.

Lingua Layer is the first product of Langdogu and its browser arm. Version 0.2.1 has been submitted to Firefox Add-ons and is in review.

What we are building next

• The ontology core. A concept layer that sits above any single language, fixed expressions whose meaning is not the sum of their characters, spaced-repetition state (FSRS) for every item, and a record of where each word actually appeared in what you read and watched. Designed; implementation in progress.
• Known-word highlighting. Sentence-level AI judgment of which words you already know, with level filters such as JLPT, so you can read at the edge of your ability. Planned.
• Model routing. Light models handle preprocessing and most sentences; idioms and ambiguous passages are escalated to stronger models. Planned.
• More surfaces. Subtitle-based video learning cut by utterance rather than by scene, and audio, all sharing the same vocabulary and review state. Planned.
• Open by design. We intend to open-source the core, keep self-hosting free, support Anki import and export, and offer an MCP server so AI assistants can save words straight into your Langdogu vocabulary. Revenue would come from a managed cloud subscription.

Langdogu is an early-stage, solo, pre-incorporation project. The ontology and learning product are in design and development; Lingua Layer is the first product. Built with Claude Code and Codex.