A textbook you can question, not just read
Static textbooks answer in chapters; students ask in sentences. The Humanoid Robotics Textbook is an interactive educational platform where the content is embedded into a vector database, so a question finds the right passage by meaning — even when the student's wording matches nothing in the index.
- Type
- EdTech · interactive textbook
- Stack
- TypeScript · Next.js · Docusaurus · Qdrant
- Search
- Semantic — embeddings over textbook content
- Status
- Live at humanoid-robotics-textbook.vercel.app
The problem
Technical curriculum lives in long-form documents, but learners arrive with specific questions — “why do bipedal robots need ankle actuation?” — that a table of contents can't route. Keyword search fails the same way: students don't know the textbook's vocabulary yet, which is precisely why they're reading it.
What it does
The textbook content is authored in Docusaurus for clean structure and rich multimedia, then chunked and embedded into Qdrant. A semantic search layer matches student questions to passages by meaning rather than exact words, turning a linear textbook into something closer to a subject-matter tutor that always cites where in the book the answer lives.
Build notes
Built with TypeScript and Next.js, with Docusaurus for the content layer and Qdrant as the vector store. The chunk-embed-retrieve pipeline here is the same foundation I use in production RAG systems — applied to education instead of compliance or support.