Reported issues for zoteus

Pod holds 8 of 8 problems reported by people outside the maintainer team. Issues filed by the project's own owners, members and collaborators are excluded entirely — a maintainer's release checklist is not a warning to a prospective user.

Back to zoteus.

Most discussed

Feature Request: Robust local fulltext extraction (any PDF/EPUB, not only Zotero-indexed attachments)

Context

I run Zoteus as my primary Zotero MCP server (HTTP transport, ~10,000 items, OpenAI text-embedding-3-large index, SQLite backend). Benchmarks and an architecture analysis on my website (https://www.logies.de/zotero-mcp-server-beaver-comparison.html) show Zoteus is clearly superior to the alternative zotero-mcp-server (54yyyu/zotero-mcp) for day-to-day operations: ~100x faster data operations (< 0.01 s vs. 2–7 s per call), better tool coverage, incremental index updates since v1.7.0,

Read the thread · 2026-08-28 · closed · external user · 1 comment

Local API connection check appears to run only once at startup, not retried later

Local API connection check appears to run only once at startup, not retried later

Summary

zotero_whoami reports localApi: false whenever Zotero's local HTTP server (port 23119) wasn't already up at the moment Claude Desktop (and the Zoteus extension) launched. Once that initial check fails, it stays false for the rest of the session, even after Zotero is confirmed running with the correct setting enabled. Restarting Claude Desktop after Zotero is already running fixes it immedia

Read the thread · 2026-08-27 · closed · external user · 1 comment

Bug: Zoteus MCP (mcpb install) crashes on startup in most recent Claude Desktop version

Claude Desktop was recently updated to include a "shared pool for Cowork and Code", i.e. shared memory, skills, and MCPs. This seems to break mcpb installs of Zoteus, so that the MCP server crashes on launch.

Here is a log:

2026-08-26T07:29:42.331Z [Zoteus — Zotero MCP] [info] Initializing server... { metadata: undefined } 2026-08-26T07:29:42.334Z [Zoteus — Zotero MCP] [info] Using built-in Node.js for MCP server: Zoteus — Zotero MCP { metadata: undefined } 2026-08-26T07:29:43.774Z

Read the thread · 2026-08-26 · closed · external user · 9 comments

Feature request: surface fulltext option variables in settings for desktop installs (mcpb)

Would it be possible to surface the fulltext_max_chars and limit variables in the desktop extension configuration menu?

Particularly since there is a default max_chars limit, it would be useful for users to actually see it and easily adjust it. Some (like me) would probably want full coverage, thus setting max_chars to 0, even if it takes days to do the initial embed. (With a different MCP tool it took close to a week on my old laptop to embed my ~4000 item Zotero library, which includes many

Read the thread · 2026-08-21 · closed · external user · 0 comments

Documentation / UX: Guide users to build the semantic-search index before first use

Summary

zotero_semantic_search returns an empty result set until zotero_index build has been run. A new user may not realize this and conclude that semantic search is broken.

Observed behavior

zotero_semantic_search → ❌ Index empty, zotero_index build needed

After running zotero_index build the feature presumably works, but the need to do so is easy to miss.

Suggested improvements

1. Actionable error message: When the index is empty, return a message such as “Sem

Read the thread · 2026-08-13 · closed · external user · 1 comment

Feature request: Fallback PDF full-text extraction when Zotero has not indexed the file

Summary

zotero_get_fulltext in Zoteus only returns content if Zotero has already indexed the PDF. For libraries where many PDFs are not yet indexed (or where indexing was skipped), this leaves a gap that other tools fill differently.

Observed behavior

zotero_get_fulltext → ❌ only for indexed PDFs

Workarounds we use today

• pyzotero with local=True reads the /fulltext endpoint very quickly (≈ 0.06 s for a 40 k-character PDF; 442-page textbook in one call).
• The olde

Read the thread · 2026-08-13 · closed · external user · 1 comment

Feature request: Support DOI import without a local translation server

Summary

During a multi-tool evaluation of Zotero-MCP servers (August 2026) we found that zotero_import in Zoteus reliably fails for DOI import unless a local Zotero translation server is running in Docker. This makes the otherwise very fast and stable MCP server harder to deploy for users who just want to add items by DOI.

Observed behavior

zotero_import (DOI) → ❌ Translation-Server needed

In contrast:

• The older zotero-mcp-server imports DOIs directly via the local Zotero

Read the thread · 2026-08-13 · closed · external user · 1 comment

update_item / create_items cannot write array-valued fields (creators, tags, collections) — Zotero rejects with "property must be an array"

Hi!

First, a sincere thank-you for Zoteus. The thoughtful feature set makes it a great tool. I really appreciate the care that has gone into it.

I think I may have run into a small bug, and I wanted to report it in case it is useful. I have not been able to find a workaround on the client side.

Summary

When the patch of zoteus_zotero_update_item (or the item body of zoteus_zotero_create_items) contains a field whose value must be a JSON array of objects — e.g. creators, tags, collections

Read the thread · 2026-07-15 · closed · external user · 1 comment

Most recent

The remaining reports are on the project's issue tracker.