Open source · CLI · Python
A CLI for building persistent knowledge bases with LLMs.
Drop in source documents. The LLM reads them, writes structured markdown pages, maintains cross-links, and keeps everything up to date — so your knowledge compounds instead of disappearing.
pip install wiki-builder
Why wiki-builder?
Not RAG
RAG re-derives context from raw documents on every query. wiki-builder integrates sources into a persistent wiki that grows smarter over time.
Plain markdown files
The knowledge base is just a folder of .md files. Open it in Obsidian, VS Code, or any editor. No database, no lock-in.
LLM does the work
The agent reads your schema, inspects the existing wiki, and writes summaries, concept pages, entity pages, and cross-links in one pass.
Quickstart
Get from zero to a working wiki in a few commands.
pip install wiki-builderexport ANTHROPIC_API_KEY=sk-ant-...
# or
wiki config --provider anthropic --api-key sk-ant-...wiki init my-research/ --name "AI Research"cp ~/papers/paper.md my-research/raw/cd my-research
wiki ingest raw/paper.mdThe LLM writes a source summary, creates concept and entity pages, and updates the index.
wiki query "what is self-attention and why does it matter?"wiki status
wiki lintCommands
wiki init <path> |
Scaffold a new wiki project with the right directory structure and a schema file. |
wiki ingest <file> |
Ingest a source document into the wiki. Supports .md, .txt, and PDFs. |
wiki query "<question>" |
Ask a question. The LLM reads the wiki and synthesizes an answer. |
wiki lint |
Health check — finds orphan pages, missing links, contradictions, and stale claims. |
wiki status |
Show page counts and word totals. No LLM call. |
wiki config |
Set default provider, model, and API key. |