Central LLM WikiCompareSign in← Home

Central LLM Wiki vs NotebookLM

Both let you drop in sources and ask questions. The difference is what is left behind afterwards.

Choose Google NotebookLM if

You want quick Q&A and audio overviews over one batch of documents, are not building something that accumulates, and do not need agents like Cursor or Claude to read and write it.

Choose Central LLM Wiki if

You want every source compiled into a linked, editable wiki page that compounds over time, and you want Cursor, Claude and ChatGPT to read the same knowledge base through MCP, with designated agents allowed to write it.

FeatureGoogle NotebookLMCentral LLM Wiki
MechanismRetrieval: each question re-reads the sources and cites passagesCompilation: the AI writes persistent wiki pages, an index and a log (Karpathy’s LLM Wiki)
What remainsThe chat and notes; the sources stay as they wereInterlinked Markdown pages you can edit, version and export
Agent accessNo public API or official MCP for the consumer versionBuilt-in MCP server: token (Reader or Librarian) or read-only OAuth, for Cursor, Claude Code, Claude.ai and ChatGPT
ModelGemini, included in the planSelf-host and configure your own model provider and API key
BibliographyCitations point to source passagesDOI/Crossref metadata, BibTeX/CSL import and export, and [@citekey] citations
ExportWhole-notebook Markdown export: not foundOne-click zip, Obsidian-compatible
DeploymentGoogle-hosted serviceSelf-host on infrastructure you control; open source under Apache-2.0
Open sourceNoApache-2.0, self-hostable

Trade-offs we admit

  • NotebookLM’s Q&A and audio overviews are excellent; if that is all you need, it is less work.
  • With Central LLM Wiki you run and maintain the deployment yourself and pay the model provider directly, an explicit trade-off.
  • Prices and features are as published on the other product’s site; the check date is in the footer.
Get started on your own serverTheir facts checked on 2026-09-04 · Google NotebookLM Website