Zotero BibTeX Sync

Syncs a .bib file to a Zotero collection — creates missing items, skips duplicates.

Setup (one-time)

Add "api_key" to ~/.config/zotero-mcp/config.json:

{
  "library_id": "...",
  "library_type": "user",
  "api_key": "YOUR_KEY_HERE"
}

Generate a key at https://www.zotero.org/settings/keys (enable library write access).

Usage

Find the collection key using the mcp__zotero__zotero_get_collections tool — each collection has a Key field.

Run the sync:

cd ~/.claude/skills/zotero-bib-sync && uv run sync.py --bib <path-to-bib> --collection <collection-key>

Dry run (preview without writing):

cd ~/.claude/skills/zotero-bib-sync && uv run sync.py --bib <path-to-bib> --collection <collection-key> --dry-run

Important: Ensure URLs are present

Before syncing, always check that each BibTeX entry has a url field. Zotero cannot fetch full text without it. For arXiv papers, add url = {https://arxiv.org/abs/<ID>} derived from the eprint field. For published papers, use the DOI link url = {https://doi.org/<DOI>}.

If items were already synced without URLs, run the add_urls.py script to backfill them:

cd ~/.claude/skills/zotero-bib-sync && uv run add_urls.py

Dedup Logic

Items are matched by: DOI → arXiv ID (from eprint field) → normalized title. Existing matches are skipped.

BibTeX Field Mapping