Overview
Open Notebook is a self-hosted, source-centered workspace for AI-assisted research. The series follows one research lifecycle: define notebook scope, preserve relationships among sources, insights, notes, and conversations, extract and index content, choose lexical, vector, Ask, or Chat retrieval, then review reusable notes and extend them through podcasts, APIs, or MCP. Its value is visible control over context, providers, and operational states—not a guarantee of truth, complete extraction, or privacy. Trust still depends on verification and operator duties such as workers, backups, credentials, and perimeter hardening.
Learning outcomes
- Open Notebook keeps source material, derived insights, notes, and conversations distinct while linking them for research reuse.
- A source becomes usable in stages; scope, extraction, indexing, and retrieval readiness can fail independently.
- Lexical search verifies exact language, vector search discovers related ideas, Ask synthesizes broadly, and Chat follows selected context.
- Transformations and notes make repeated analysis reusable, but important claims still need review against their source trail.
- Self-hosting, podcasts, and API or MCP access extend the workflow, but privacy, permissions, and factual trust remain conditional.
Keywords
Preview every chapter · Unlock full text, audio, and video from chapter 4
PlansFAQ
What does Open Notebook keep separate in a research workflow?
Open Notebook separates sources, derived insights, editable notes, and conversations while linking them through notebook scope.
Why can a source appear in Open Notebook but fail to support search?
A source can appear before extraction, chunking, or embedding is complete, or it can fall outside the selected scope; readiness is not trust.
Which Open Notebook retrieval mode fits an exact or broad question?
Use lexical search for exact words, numbers, and quotations; vector search for related ideas; Ask for broad synthesis; and Chat for selected context and follow-up.
How do Open Notebook transformations and notes support reusable research?
Transformations apply a repeatable analytical question to source content, while notes preserve reviewed conclusions, comparisons, and unresolved points for later reuse.
Does self-hosting make Open Notebook fully private?
No. Self-hosting can keep storage and selected processing under the operator's control, but hosted providers, network services, credentials, workers, and configuration still define the actual boundary.
Can Open Notebook podcasts or API and MCP outputs replace source verification?
No. Podcasts and client-generated results are transformations or access paths over the same research graph, so inspect transcripts, scope, citations, and source passages before relying on them.