Overview
PRIMA Scholar is a research workspace for Claude Code, and this page describes version 2.2.1. It searches ten academic databases from one request and keeps your papers in a library on your own computer. It also helps you write with citations in APA7 or five other styles.
PRIMA Scholar is a standalone product with no dependency on AI Business OS or any other PRIMA product. You can install it on its own in any Claude Code setup.
What It Does
Section titled “What It Does”- Academic search: one search reaches all ten databases, and when several databases return the same DOI it keeps the copy with the most complete details.
- Search wizard: the wizard reads your query, detects its discipline and suggests the databases that suit that field.
- Open access first: every result shows whether the paper is open access or gated, with open access papers listed first by default.
- Full text and download links: PRIMA Scholar returns full text from CORE when you have a CORE key, or a download link from Europe PMC, arXiv or bioRxiv/medRxiv.
- Research library: you can import PDFs, Word documents, Markdown and text files into a local database with full-text search.
- Collections and tags: you can group library records into named collections and label them with tags.
- Citations in six styles: search results carry ready-made citations in APA7, Harvard, Chicago, Vancouver, IEEE and MLA.
- Citation tracking: you can follow a paper forward to the work that cites it, or backward to its references.
- Writing with citations: Claude drafts prose with in-text citations and a reference list drawn from your sources.
- Literature watch: a recurring scan of your standing topics reports only the papers published since the last run.
- Research agents: three agents handle longer jobs, from a full literature synthesis to filling in a structured form.
How It Is Built
Section titled “How It Is Built”PRIMA Scholar has two parts that work together.
The first part is two MCP servers, which are small programs that run on your computer and give Claude new tools. One of them, prima-scholar-search, talks to the ten databases. The other, prima-scholar-library, keeps your research library in a single database file. The Tools Reference page describes every tool the two servers provide.
The second part is a set of commands, skills and agents that guide Claude in using those tools. These run only in Claude Code.
| Part | What You Get |
|---|---|
| Commands | /scholar, /cite, /library and /literature-watch |
| Skills | researching-topics, managing-research-library, writing-with-citations and literature-watch |
| Agents | research-agent, form-completion-agent and literature-watch-agent |
| MCP servers | prima-scholar-search (five tools) and prima-scholar-library (twelve tools) |
A Claude Code plugin install adds the plugin name in front of each command, so /scholar becomes /prima-scholar:scholar. The pages in this section use the short names for readability.
Where It Runs
Section titled “Where It Runs”The full product runs in Claude Code, and Installing in Claude Code walks you through it.
What You Need
Section titled “What You Need”- Claude Code, on a Claude plan that includes it.
- Node.js 18 or later and npm, to build and run the two MCP servers.
git, to download and update PRIMA Scholar.
No API key is needed to start, because nine of the ten databases work without one. CORE needs a free key, and every other setting is optional. The details are on API Keys and Configuration.
Where to Go Next
Section titled “Where to Go Next”- Follow Installing in Claude Code to install PRIMA Scholar and build its two servers.
- Follow Searching the Literature to run your first search across the databases.
- Follow Your Research Library to save the papers and documents you want to keep.
- Read Citing Sources in APA7 to see how citations reach your writing.