Zotero MCP is a feature-rich server that seamlessly integrates your Zotero library with AI assistants. It offers core capabilities like search, semantic search, annotation extraction, and item management, with optional extensions for scite citation intelligence and advanced PDF handling. Setup is straightforward via CLI or manual config, and the documentation is comprehensive. The main risks involve API key handling and external service dependencies, which users should manage carefully. Overall, it is highly suitable for researchers and students.
Strongest · Documentation 16/20 Weakest · Reliability 11/20
Static review · not runListed 2026-08-07
Read the FMRS scoring method →Runtime：Python 3.10+ · Docker
ZOTERO_API_KEY
Zotero Web API key from https://www.zotero.org/settings/keys; needed when ZOTERO_LOCAL is false or for writes in hybrid mode.
OPENAI_API_KEY
OpenAI API key (platform.openai.com) required for OpenAI embeddings.
GEMINI_API_KEY
Google Gemini API key required for Gemini embeddings.
ZOTERO_WEBDAV_PASSWORD
Optional WebDAV password.
ZOTERO_LOCAL
Set to true to read from the local Zotero desktop app; requires Zotero 7+ with the local API enabled in settings.
ZOTERO_LIBRARY_ID
Zotero user or group library ID; your numeric userID is shown on the same API keys page.
ZOTERO_LIBRARY_TYPE
Whether the library ID is a personal (user) or group library; defaults to user.
ZOTERO_MCP_TOOLSETS
Comma-separated list of optional tool groups to expose (all/none/named groups); controls the tool surface.
ZOTERO_EMBEDDING_MODEL
Embedding provider for semantic search: default, openai, gemini, or ollama.
OPENAI_EMBEDDING_MODEL
OpenAI embedding model name, e.g. text-embedding-3-small or text-embedding-3-large.
OPENAI_BASE_URL
Optional custom OpenAI-compatible endpoint URL.
GEMINI_EMBEDDING_MODEL
Gemini embedding model name, e.g. gemini-embedding-001.
GEMINI_BASE_URL
Optional custom Gemini-compatible endpoint URL.
OLLAMA_EMBEDDING_MODEL
Model name for local Ollama embeddings, e.g. nomic-embed-text or bge-m3.
OLLAMA_BASE_URL
Ollama server URL, default http://localhost:11434.
ZOTERO_DB_PATH
Custom path to zotero.sqlite; auto-detected when unset.
ZOTERO_WEBDAV_URL
Optional WebDAV folder URL for direct attachment downloads in remote mode.
ZOTERO_WEBDAV_USERNAME
Optional WebDAV username.
ZOTERO_MCP_SCHEMA_REFRESH
Set to 0 to disable the weekly background refresh of Zotero's item-type schema.
ZOTERO_MCP_SCHEMA_CACHE
Custom path for the schema cache, default ~/.cache/zotero-mcp/schema..
ZOTERO_APP
Docker container runtime mode: server (default) or cli.
ZOTERO_TRANSPORT
Transport method inside the Docker container, default stdio.
zotero-mcp setup for auto-configuration (supports Claude Desktop), or manually add the JSON config to your MCP client as shown in the docs.{
"mcpServers": {
"zotero": {
"command": "zotero-mcp",
"env": {
"ZOTERO_LOCAL": "true",
"ZOTERO_API_KEY": "YOUR_API_KEY",
"ZOTERO_LIBRARY_ID": "YOUR_LIBRARY_ID"
}
}
}
}
Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.
{
"servers": {
"zotero": {
"command": "zotero-mcp",
"env": {
"ZOTERO_LOCAL": "true",
"ZOTERO_API_KEY": "YOUR_API_KEY",
"ZOTERO_LIBRARY_ID": "YOUR_LIBRARY_ID"
}
}
}
}
Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).
claude mcp add zotero -e ZOTERO_LOCAL=true -e ZOTERO_API_KEY=YOUR_API_KEY -e ZOTERO_LIBRARY_ID=YOUR_LIBRARY_ID -- zotero-mcp
Run it in a terminal; replace any <…> placeholders with your own values first.
After running zotero-mcp setup, restart your client and check that zotero_ tools such as zotero_search_items and zotero_get_item_metadata appear in the client's tool list; a prompt like "Search my library for papers on machine learning" returning real results confirms the connection.
Once connected, you can ask your AI assistant things like:
No matching tools
Listed from the project's documentation, not tested by this site.
Zotero MCP connects your Zotero research library with AI assistants via the Model Context Protocol. It supports both local and web access, offering AI-powered semantic search (optional extra), paper search, summaries, citation analysis, PDF annotation extraction, and write operations like adding papers by DOI or URL, managing collections, and merging duplicates. Requires Zotero 7+ with local API enabled or a Zotero Web API key for write operations. Optional tool groups (e.g., scite, duplicates) can be enabled via environment variables.
Manage your Hevy workout data from AI assistants via MCP.
Firecrawl's official MCP server for web search, scraping, and structured extraction for AI agents
One binary. One grammar. Evidence from the biomedical sources you already trust.
Encrypted, fully offline long-term memory for AI agents
Source revision eb016b1b746d Data synced 2026-10-19 Read the FMRS scoring method