Analysis updated 2026-05-18
Verify every citation in a research draft against scholarly databases before finalizing it.
Catch a citation to a paper that has since been retracted.
Check that a citation's title, author, and year all match the actual source.
Run an autonomous research job on a topic from the terminal.
| vibhorxpandey/aurelius | 0xustaz/streamgate | a-bissell/unleash-lite | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | Python | Python | Python |
| Setup difficulty | easy | hard | hard |
| Complexity | 2/5 | 4/5 | 4/5 |
| Audience | researcher | developer | researcher |
Figures from each repo's GitHub metadata at analysis time.
Optional Tavily API key for general web search, autonomous mode needs a separate LLM API key.
Aurelius is a research assistant server built on the Model Context Protocol, a standard that lets AI chat apps like Claude, Gemini CLI, Cursor, and others plug in extra tools. Aurelius adds a set of research tools that check every citation and claim a language model produces against real scholarly databases and live web sources, so the app can catch made up papers or citations to studies that have since been retracted before that content ever reaches the final draft. By default Aurelius runs in what the README calls host driven mode: the chat app you connect it to uses its own subscription and its own model to think and write, while Aurelius only supplies the fact checking tools. This means Aurelius itself needs no separate AI subscription to run in this mode. Citation checking runs against two free scholarly databases, OpenAlex and Crossref, which need no account. An optional key for a service called Tavily can be added to enable general web search and to act as a backup when a citation cannot be found in either scholarly database. There is also a separate autonomous mode where Aurelius drives its own research using its own AI model, which does require an API key with its own usage quota. The README walks through real examples of what the tool catches: a citation with the wrong author attached to a correct sounding title, a paper that has since been formally retracted, and a citation with a title that matches but points to an entirely different paper once the author and year are checked. In each case, the tool does not just confirm that a title exists somewhere, it cross checks the author, year, and retraction status before calling a citation verified, and it hands back a corrected citation when it finds a better match. Setup is a standard Python package install, after which you register Aurelius with your chat app of choice by adding a small configuration snippet, then simply ask the app to use Aurelius to research a topic and verify the citations. The project provides individual tools for screening a topic, drafting an outline, verifying single citations, verifying whole batches of claims, and verifying an entire references section at once. The README notes the project is still under active development. It is released under the MIT license.
An MCP research server that plugs into Claude, Gemini CLI, and Cursor to fact-check citations and claims against real scholarly databases before they reach a draft.
Mainly Python. The stack also includes Python, MCP, OpenAlex.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly researcher.
This repo across BitVibe Labs
Verify against the repo before relying on details.