cyberlibrary_search for a unified BM25 + semantic query, cyberlibrary_fulltext_search for literal text search, cyberlibrary_get_chapter for clean Markdown and cyberlibrary_cite for ready citations. It shares a 5-calls/day free bucket per IP; past that, 1–5 credits over x402.CyberLibrary is a public-domain corpus of 80,354 books in 23 languages, served as chapter-level Markdown plus a searchable paragraph index. It is not a generic web search: every result carries a stable book/chapter/paragraph identifier, a citation and a license/source URL so an agent can quote it safely.
| Tool | What it does |
|---|---|
cyberlibrary_search | Unified search: automatic language detection, BM25 + neural semantic fusion (RRF), ranked passages. |
cyberlibrary_fulltext_search | Literal full-text search inside the actual book text; returns the matching passage, book, chapter and canonical URL. |
cyberlibrary_search_passages | BM25 passage retrieval with chunk_id, citation and license. |
cyberlibrary_semantic_search | Neural embedding search for “describe what you are looking for”. |
cyberlibrary_rag_context | Packs ranked passages into a token budget, ready for agent context. |
cyberlibrary_book_meta / get_toc | Book metadata and ordered chapter titles. |
cyberlibrary_get_chapter / get_markdown | Clean Markdown of one chapter, or the whole-book mirror. |
cyberlibrary_cite / verify_quote | Ready-to-paste citation; verify whether a quotation actually occurs in the corpus. |
cyberlibrary_get_chunks | Stable paragraph-level chunks with permanent chunk_id. |
cyberlibrary_catalog_index / language_shard | Language catalogs with counts; paginated catalog pages. |
cyberlibrary_service_info | Discovery endpoints for the library (MCP, catalog API, llms.txt, OpenAPI, HTML UI). |
Search 19th-century Russian novels for passages about a duel, with citations. Find every mention of "ledger" in Moby Dick and give exact chapter citations. Give me a 2,000-token RAG context from public-domain books about lighthouse keepers. Verify this quotation: "..." — which book and chapter is it from?
{
"mcpServers": {
"robotbase": { "type": "http", "url": "https://robotbase.cc/mcp" }
}
}
16 CyberLibrary tools live behind the same endpoint as web search and on-chain verification. No key, no account.
CyberLibrary shares the 5-calls/day free bucket per IP with the two TRON risk tools. Past that, retrieval costs 1–2 credits and full-text search 5 credits, paid over x402. The corpus is public domain only — no current copyrighted books — and every passage returns its license and canonical source URL. This is not legal advice; check the license/territory notice attached to each book.