Nodus: an open source research workspace for academic work
Academic research now runs on a startling number of applications. References in Zotero. Books and papers as PDFs in a folder. Notes in something else. A spreadsheet doing the job of a database. And, for the last couple of years, an AI assistant in a browser tab that knows nothing about any of it.
Each of those tools is good at its job. The trouble is the space between them, which is where a surprising share of the working day goes: copying a citation, hunting for the PDF that goes with it, pasting a passage into a chat window, then losing the answer when the tab closes.
Nodus is an open source research workspace built to close some of that distance. It holds your documents and references, your notes and ideas, structured research databases, and AI that works on your own sources. It can run its own library or read an existing Zotero one, so adopting it doesn't mean abandoning what already works.
The aim is a practical place to do the ordinary daily work of researchers, students and teachers. A scholar's workspace, in other words, rather than one more container for files.
A research workflow built around sources, ideas and connections
The premise is unremarkable once you say it out loud. Research isn't just collecting documents. Sources produce ideas, ideas connect to other ideas, arguments grow out of those connections, and at the end every sentence has to be defensible against a specific piece of evidence. Software that treats every document, note and reference as an unrelated item throws away exactly the part that took you longest to build.
So Nodus keeps a library where documents and bibliographic data sit alongside notes, ideas and the rest of the project. You read documents in the app, search them semantically, and connect the ideas pulled from a source to everything else. You can move from one document to the claims inside it, then out to where those claims are being used.
There are also tools for comparing across the whole corpus: finding relations between ideas, laying out agreements and contradictions, and spotting where the material you have doesn't really answer the question you're asking. These assist. They don't adjudicate, and the interpretation stays yours.
All of this matters more the longer a project runs. A thesis can accumulate several hundred books and articles, thousands of pages of notes and a growing tangle of arguments and references. Year one you can hold it in your head. Year three you cannot, and whether the connections survive is down to whether anything wrote them down.
AI in the research process rather than beside it
Nodus includes a fair amount of AI, but the goal was never to bolt on another general-purpose chatbot. Everything is pointed at the material you're actually working with.
Documents get analysed for ideas and the passages that support them. Semantic search finds related material across a collection even when the vocabulary doesn't match. The library can feed conversations and longer research reports, and where citations exist, generated text stays connected to the sources it came from.
You choose the provider, and you can point Nodus at a compatible local model instead. That keeps control over how your material is processed and stops the whole workspace from depending on one company's pricing decisions.
This distinction is worth being blunt about. Academic work needs more than a plausible answer, and a plausible answer with no traceable source is worse than no answer, because it's harder to catch. You need to know what evidence stands behind a claim and be able to open it. AI can find and arrange things very well. Deciding what it means is still the job.
Vaults for different kinds of research and learning
Projects don't all want the same shape, which is why Nodus splits into vaults. Each one is a workspace tuned to a kind of work, sitting on the same engine for documents, information and knowledge.
The Academic vault handles research projects built on sources, references, notes, ideas, databases and analysis. Teaching covers university and school workflows: materials, courses, assessments, reusable resources. Study is built around learning, revision and course material. Databases gives you a flexible home for projects that have become structured data.
Other vaults handle more specialised work, including genealogy, worldbuilding, primary sources and oral history. Nobody is expected to use all of them. Open the one that matches your project and ignore the rest; they don't clutter the interface of a vault you aren't in.
Nodus also talks to tools you probably already have. The Zotero integration puts an existing reference library into the research workflow. A browser extension collects material from the web. Word and LibreOffice integrations bring research material into the writing itself, rather than leaving you to retype it.
An open source project that came out of academic research
Nodus started as a personal project. It exists because I needed to solve some very specific, very tedious problems while organising material for a PhD, and the tools I wanted didn't exist in a form I could use. It grew from there into a general research and knowledge workspace, mostly by adding whatever turned out to be missing next.
That origin still shows in how it develops. Features get tested against real research work instead of a list of productivity features that sound good on a landing page. Some of the results are slightly odd if you're coming from consumer software, and I think that's usually correct: research is odd. What I care about is making large collections of sources workable, and keeping the material underneath accessible to the person doing the research.
Nodus is free and open source under AGPL-3.0-only. It's local-first, and the desktop application needs no account. Your documents stay yours, and you decide which external services, if any, ever see them. The open-source and local-first guide sets out the licence, the storage model and every outgoing request.
It's still very much under development, and plenty of the workflow could be better. The long-term goal was never to replace every application a researcher uses. It's to give the sources, notes, ideas and daily activity of a project one place where they stay connected for as long as the project lasts.
Explore Nodus for academic research