Get AI summaries of any video or article — Sign up free
Zettelkasten workflow for research papers | Zotero & Obsidian link thumbnail

Zettelkasten workflow for research papers | Zotero & Obsidian link

Artem Kirsanov·
5 min read

Based on Artem Kirsanov's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.

TL;DR

Use Zotero as a centralized intake hub for research sources, relying on quick add (DOI/ISBN), web clipping, and optional Research Rabbit discovery to reduce manual entry.

Briefing

A research workflow built around Zotero plus Obsidian turns scattered PDFs and citations into a connected system of “literature notes” that can later become atomic Zettelkasten-style idea nodes. The core payoff is staying oriented in a growing sea of sources: every paper gets a stable reference entry, and new notes can link back to it automatically—so later research doesn’t require hunting through Zotero or re-copying bibliographic details.

Zotero serves as the first hub for collecting papers and books. Instead of relying on desktop folders, the workflow uses Zotero’s quick add (paste a DOI or ISBN to auto-fill metadata and attach PDFs when available) and a Zotero connector for clipping from web pages. A further automation option links Zotero with Research Rabbit so newly discovered papers can land directly in the Zotero library. Once papers accumulate, organization happens through two complementary structures: subcollections (used like nested folders) and tags. Subcollections group sources by context—such as lab research projects or materials gathered for a presentation—while tags capture topic-level meaning in a flexible, free-form way (for example, tagging a paper by hippocampus place cells remapping and optogenetics). Color-coded tags act as quick status markers, such as “already read” or “high priority.”

To bring this structure into Obsidian, the workflow relies on two plugins: Better BibTeX for exporting Zotero metadata and Citations for importing it into Obsidian. Better BibTeX exports the Zotero library as a continuously updated JSON file (using “keep updated”), and Citations reads that file to generate reference-backed notes inside a dedicated “literature notes” folder. When writing in Obsidian, inserting a literature note is fast: a command lets the user search for the target Zotero item and create (or reuse) a markdown note tied to a unique “site key.” Clicking that note reveals the bibliographic metadata and provides a navigation anchor for the paper.

Those literature notes do more than store citations. They can remain empty as organizational placeholders, but they also function as a reading journal. As ideas are extracted from a paper, the notes can hold outlines of key points, terminology, author moves, and personal remarks—especially useful because an “atomic idea” may span multiple pages and detours that are hard to track while reading. The workflow also encourages delayed promotion: literature notes don’t have to become Zettelkasten nodes immediately. Sometimes the connections only crystallize months later when a related source is encountered, at which point the existing reference note provides the groundwork for rewriting a clearer, more atomic idea.

After reading, optional “post processing” deepens retention and understanding. The workflow suggests spaced repetition via flashcards created inside Obsidian using an Anki integration plugin, with questions and answers stored in Obsidian note containers. For deeper comprehension, it also supports embedding mind maps (often drawn on an iPad) into the literature note as images, allowing quick refresh when revisiting the source. The overall message is pragmatic and non-rigid: build the reference backbone first, capture ideas as they emerge, and only convert them into atomic nodes when the time is right.

Cornell Notes

Zotero organizes research sources with metadata automation (quick add, web clipping, and optional Research Rabbit discovery) and a two-layer system: subcollections for context and tags for topic meaning and status. Better BibTeX exports the Zotero library as an always-updating JSON file, and the Citations plugin in Obsidian uses that file to create “literature notes” linked to each Zotero item via a unique site key. These literature notes can act as placeholders, but they’re most useful as a reading journal—outlining key points, terminology, and personal remarks while extracting ideas from the paper. Promotion into atomic Zettelkasten nodes can be delayed until connections become clear, sometimes months later. Optional post-processing includes spaced-repetition flashcards (via Obsidian-to-Anki) and embedded mind maps for fast review.

Why use both subcollections and tags in Zotero instead of only folders?

Subcollections (nested like folders) group papers by context where knowledge will be applied—such as lab research projects or sources collected for a presentation. Tags handle topic-level classification in a flexible, free-form way based on what the abstract suggests (e.g., hippocampus place cells remapping plus optogenetics). This lets the library be filtered by tags to find all papers about a topic, while subcollections keep workstreams and use-cases separate. Color-coded tags add a lightweight status system like “already read” or “high priority.”

How does the workflow connect Zotero items to Obsidian notes without manual retyping?

Better BibTeX exports the Zotero library metadata into a JSON file (set to “keep updated”), including titles, authors, years, and URLs. In Obsidian, the Citations plugin is configured with the path to that JSON file. When inserting a literature note, Obsidian searches the target Zotero item and creates a markdown note populated from the metadata. Referencing the same article again links to the existing literature note instead of creating duplicates, using a unique “site key” as the identifier.

What role do literature notes play in a Zettelkasten-style research pipeline?

Literature notes can be empty placeholders that serve as stable anchors for later linking. More importantly, they act as a reading journal: while reading, the user outlines key points, terminology, and author-specific moves, plus personal remarks. This matters because an “atomic idea” may be scattered across many pages, with examples and detours that are easy to forget mid-read. The literature note preserves that structure until the ideas are ready to be rewritten as atomic nodes.

Why delay converting literature notes into atomic Zettelkasten nodes?

Some ideas take multiple chapters to fully develop, and even after finishing a paper, the connections needed for a clean atomic formulation may not be available yet. The workflow treats this as normal: months later, encountering a related source can provide the missing context. At that point, the existing literature note can be revisited and combined with fresh details to reach “critical mass” for a sophisticated, simple, atomic rewrite.

What post-processing steps help turn reading into long-term recall?

Two main options are suggested. First, spaced repetition: flashcards can be created inside Obsidian using an Obsidian-to-Anki plugin, with questions and answers stored in Obsidian note containers (the “language notes” mentioned). Second, mind maps: drawing a conceptual map (often on an iPad) and embedding it as an image inside the literature note lets the user refresh memory quickly when revisiting the paper. Because mind maps are time-consuming, they’re reserved for especially relevant papers.

Review Questions

  1. How do subcollections and tags complement each other in Zotero, and what kinds of decisions does each support?
  2. Describe the chain from Zotero metadata export to Obsidian literature note creation, including the purpose of the JSON file and the site key.
  3. What criteria would make you convert a literature note into an atomic Zettelkasten node immediately versus waiting until later?

Key Points

  1. 1

    Use Zotero as a centralized intake hub for research sources, relying on quick add (DOI/ISBN), web clipping, and optional Research Rabbit discovery to reduce manual entry.

  2. 2

    Organize Zotero libraries with subcollections for context (projects, presentations) and tags for topic meaning and reading status.

  3. 3

    Export Zotero metadata with Better BibTeX into a continuously updated JSON file, then configure Obsidian’s Citations plugin to read that file.

  4. 4

    Create “literature notes” in Obsidian via fast search and insertion; reuse existing notes for the same article using the site key to avoid duplicates.

  5. 5

    Treat literature notes as a reading journal: outline key points, terminology, and personal remarks while extracting ideas from papers.

  6. 6

    Promote ideas to atomic Zettelkasten nodes only when the connections are strong enough—sometimes months later—rather than forcing immediate conversion.

  7. 7

    Use optional post-processing like Obsidian-to-Anki flashcards for spaced repetition and embedded mind maps for rapid conceptual refresh.

Highlights

Zotero subcollections handle “where this paper fits,” while tags handle “what this paper is about,” and color-coded tags add a practical read-status layer.
Better BibTeX + Citations turns Zotero metadata into Obsidian literature notes automatically, with a unique site key that prevents duplicate notes for the same article.
Literature notes can stay empty as anchors, but they shine as a reading journal that captures scattered ideas across pages until they’re ready for atomic rewrites.
Delayed promotion is built into the workflow: connections often crystallize only after encountering related sources later.
Post-processing options—spaced repetition flashcards and embedded mind maps—convert reading into recall and faster future review.

Topics

Mentioned