Docs/What is a thoughtbase?

What is a thoughtbase?

A thoughtbase is where your thinking on a subject lives and grows. Instead of scattered documents and lost chat threads, you gather what you're working through as notes — and notes link to each other, so ideas connect the way they do in your head. Some you write yourself; some you draft side by side with the AI. Over time a thoughtbase becomes a living map of what you know, not a folder of dead files.

Each thoughtbase is self-contained: one folder of plain files, with its own knowledge graph built automatically from what's inside. You can keep several — one per subject — and open whichever you're working in. Everything is stored locally, in plain text, and stays yours.

What you might build one of

A thoughtbase suits any subject you return to and want to build up over time rather than finish and forget. A few shapes it commonly takes:

A research project
A literature review, a thesis, or an ongoing investigation — papers and web pages brought in as sources, key claims extracted, and your own synthesis linking them together.
A writing project
A book, a report, or a long essay — the manuscript alongside the background notes, characters, arguments, and references that feed it.
A subject you're learning
A course or a field you're getting to grips with — explanations in your own words, flashcards, and worked examples that build on each other.
A decision or investigation
Options, evidence, and arguments laid out and connected — support, rebut, and depends-on links making the structure of the reasoning visible.
A domain you're mapping
A product area, a system's architecture, a body of law, a company's history — a reference map of what's true and how the parts relate.
A data analysis
A question you're answering with data — CSV tables you query with SQL, notes that compute, and charts that render live against the numbers.
A personal knowledge base
A durable place for what you know and keep coming back to — a "second brain" you can actually query, not just search.

What's inside a thoughtbase

A thoughtbase collects several kinds of thing. Most of it is notes; the rest is the connective tissue and structure that turns a pile of notes into something you can query and build on.

🧵

Notes

The main content — prose, tables, images, math, diagrams, and more. One note per idea, page, or piece.

 🔗

Links

Connections between notes — plain wiki-links plus typed links like supports, rebuts, and depends-on.

 📥

Sources

Outside material brought in — web pages, PDFs, and references you cite and mine for claims.

 🏷️

Properties & tags

Structured details on a note — tags, status, and other metadata that the graph reads.

 📋

Data & tables

CSV files as typed tables, and tables inside notes — queryable data that charts and cells read.

 🕸️

The knowledge graph

A precise map of titles, tags, links, and facts, built automatically on every save — queryable with SPARQL and SQL.

 💬

Conversations & proposals

Your threads with the AI, and the changes it proposes — filed as notes and claims only after you approve them.

 🧠

Thinking-tool output

Results of the Learning, Research, and Analysis skills — each returned as a note you review and keep or discard.

Where a thoughtbase lives

A thoughtbase is just a folder on your disk. Your notes are ordinary markdown files inside it, organized into subfolders however you like — readable and portable, with or without Minerva. Alongside them, a hidden .minerva folder holds the knowledge graph and other generated data, rebuilt from your files as you work.

Because it's local-first and plain-text, a thoughtbase is yours: back it up, sync it, or put it under version control like any other folder. Nothing lives in a proprietary database you can't get out of. When you're ready to share, you can publish or export a thoughtbase without giving up the original.

Creating and managing thoughtbases

You create, open, switch between, and close thoughtbases from the File menu. See Thoughtbase operations for each command and its keyboard shortcut.