zotero-mcp

Multi-strategy semantic and keyword search across Zotero libraries locates papers by combining semantic matching, keyword queries, tags, full-text, and annotation searches. The Skill applies environment-aware search strategies—batched for Claude Desktop, autonomous for Claude Code—then outputs results as markdown bibliographies with Chinese-English translations for Logseq integration.

Yes. Autonomous agent-driven multi-step searches systematically explore your Zotero collection using iterative refinement across multiple search angles. The Skill detects your environment and adapts its approach, producing formatted markdown outlines ready for knowledge management systems.

The Skill works with both. Claude Desktop performs safe batched searches; Claude Code enables autonomous agents to conduct multi-step searches independently. Environment detection automatically selects the appropriate strategy and search depth for each platform.

Plain markdown outlines with bilingual Chinese-English translations optimized for Logseq. Results include bibliographic metadata and structured formatting suitable for direct import into knowledge management and research organization systems.

Comprehensive discovery combines semantic searches, keyword queries, tag-based filtering, full-text indexing, and annotation matching. This multi-angle approach surfaces relevant papers you might miss with single-strategy search alone.