Library & sources
How the project library works — the references and data your agents read and cite from, and how to add and import sources.
The library is your project’s collection of sources and data. It’s what agents cite from, what Claim Check verifies against, and what the Full Draft Writer writes from — so the stronger your library, the better the AI’s work.
The project library
References, PDFs, notes, and data files live with the project, not in a giant unscoped thread. Because the library is project-scoped, agents see exactly the right sources for the work in front of you.
Adding sources
Add sources directly, or discover them with Sources Hub, which mines references from papers you already have and searches the web for gaps — one click to add.
How agents use it
You don’t attach sources to each request. Anything in the library is available to every agent automatically: the Full Draft Writer pulls relevant sources from it while drafting, Claim Check verifies your citations against it, and Sources Hub adds to it. Your data and result files in the project are readable too — see files & folders.
Importing references
Already manage references elsewhere? Import them from BibTeX, RIS, and CSL-JSON files straight into the library, so you don’t start from scratch.
Connectors
Multilo is building connectors to the tools researchers already use — reference managers like Zotero, and scholarly search across arXiv and OpenAlex — to bring your external world into the project. Reference-file import is available today; live connectors are rolling out.
Curate before you draft
A few minutes adding the right sources before a big writing run pays off in more grounded, better-cited output.