Mendeley with AI: Literature Reviews Need Reference Workflows, Not Just Search

Most "AI for research papers" demos stop at search: find a paper, summarize it, maybe extract a citation.

That is useful, but it is not the workflow researchers live in. A real literature review has saved papers, citation metadata, missing fields, PDF attachments, folders, group libraries, annotations, duplicates, trash, and the constant question of what should be read next.

We rebuilt the Mendeley MCP server around that full reference-management shape. An agent can now work with the library as a library, not just as a list of search results.

What an agent can do now

The Mendeley server now covers the main research-library lifecycle:

• Search personal library documents.
• Search the global Mendeley catalog by topic, DOI, identifier, title, or author.
• Create or update library document records from discovered metadata.
• Create, rename, move, and delete folders.
• Add documents to folders and remove folder membership.
• Upload PDF files to existing documents.
• List file records, download file content as base64, and delete attachments.
• List, create, update, and delete annotations.
• List, create, update, and delete private or invite-only groups.
• Move documents to trash, restore them, or permanently delete them from trash.

That combination matters because a citation record, a PDF file, a folder, an annotation, and a group are different objects. Good agent behavior depends on knowing those boundaries.

The useful workflow: from discovery to a clean library

Start with a broad research question:

Find papers about retrieval augmented generation evaluation, separate papers already in my Mendeley library from catalog results I have not saved, and build a shortlist for a literature review.

The agent can handle this in stages:

1. Search the existing Mendeley library first.
2. Use catalog search for papers that are not already saved.
3. Create library documents from the best catalog records.
4. Group saved papers by theme, method, source, or year.
5. Create a project folder and add the selected documents.
6. Identify missing citation fields before the writing phase.

That is different from "search the web for papers." It respects the user's Mendeley account as the system of record.

PDFs are attachments, not citations

One of the important fixes is file handling. In Mendeley, document metadata and uploaded files are separate. A document can exist without a PDF. A PDF can be deleted without deleting the citation record.

That lets an agent answer questions like:

For the papers in my systematic review folder, list which documents have attached PDFs, which are missing files, and which citation records look complete enough to export.

It can then upload a PDF to a known document, confirm the stored filename and MIME type, or remove a bad attachment without trashing the paper itself.

Groups make this collaborative

Mendeley groups are shared libraries, not just labels. The server now lets an agent inspect group documents and folders, and create or update private or invite-only groups where Mendeley allows it.

That unlocks team workflows:

List the newest documents in our shared Mendeley group, identify which papers need follow-up, and separate missing citation metadata from missing PDF attachments.

For a lab, research ops team, or course reading group, this is the difference between "show me papers" and "tell me what the group library needs."

Cleanup has to be explicit

Research libraries accumulate test records, old attachments, and half-imported papers. Cleanup is useful, but it has to be careful.

The server separates the destructive steps:

• Delete a file attachment when the PDF is wrong but the citation should stay.
• Trash a document when the citation record should leave the library.
• Restore a trashed document when the deletion was premature.
• Permanently delete from trash only after review.

That separation gives agents a safer path for cleanup prompts. The model does not have to guess whether "delete this" means remove a PDF, remove a folder membership, trash a citation, or permanently purge the record.

What to try first

Start with a read-only review:

Search my Mendeley library for papers about AI evaluation, group the strongest matches by method, and list the citation details I need for a literature review.

Then try a library-organization task:

Create a folder for my RAG evaluation review, add the best matching papers from my library, and tell me which papers are missing DOI or publication-year metadata.

For shared work:

List the newest papers in my shared Mendeley group and summarize which ones need team follow-up, missing metadata, or attached PDFs.

The product page has the durable reference and live example prompts: Mendeley MCP server.

FAQ

Can an AI agent search both my library and the global Mendeley catalog?

Yes. Library search looks at documents saved in the connected account. Catalog search looks for global bibliographic records that may not be saved yet.

Can it manage PDFs?

Yes. The server can upload PDFs to existing documents, list file metadata, download file content as base64, and delete file attachments.

Can it work with shared research groups?

Yes. It can list groups, inspect shared-library documents and folders, and create or update private and invite-only groups where Mendeley supports that.

Does permanent deletion happen automatically?

No. The safe workflow is explicit: remove attachments when needed, trash documents first, inspect trash, then permanently delete only when the record should not be recoverable.