GutenbergKG¶
The Knowledge Press -- a local, source-grounded way to ask better questions of great books.
Eric G. Suchanek, PhD -- Flux-Frontiers
GutenbergKG turns public-domain texts into a queryable library of knowledge graphs. Instead of searching for a word in one book at a time, you can explore an idea across works -- then read the passages that support what you find. Retrieval is local and works without an LLM; an optional local or cloud model can turn the retrieved passages into a cited synthesis, but the evidence remains visible and inspectable.
The included corpus contains 253 texts in 20 genres: literature, philosophy, drama, diaries, sacred texts, travel, natural history, and technical reference.
Where to start¶
| Installation | CLI, Docker, and Apple container setup |
| Cheatsheet | Every gutenkg command, quick reference |
| Chat UI | The Streamlit reading-room interface |
| Moots | Two books in conversation: gutenkg converse, and playing moots in the forest |
| Knowledge Press apps | The web forest (live) and the iPhone, iPad and Mac app (in development for App Store deployment) |
Corpus¶
| Books in the corpus | The full title-level catalog |
| Corpus wishlist | Planned additions, organized by genre |
| Download pipeline | How texts are fetched and catalogued |
| Ingestion pipeline | Retrieval and build architecture |
| Selective bundles | Ship a named, versioned subset instead of the full corpus |
Visualization¶
| Ray-tracing the knowledge tree | Growing a book into a 3-D tree, cast to Looking Glass or rendered analytically |
See also RunPod deployment and Partners.
The API reference is generated from the package's own docstrings.
Source¶
GutenbergKG is Elastic-2.0 and lives at github.com/Flux-Frontiers/gutenberg_kg.