cite-rag-mcp
A local MCP server for evidence-safe academic citation workflows, including verified-reference retrieval, Zotero duplicate detection and import, and Word export via Pandoc with zotero.lua.
README
cite-rag-mcp
A local MCP server for evidence-safe academic citation workflows:
- verified-reference retrieval
- Zotero duplicate detection and DOI import
- Better BibTeX citekey extraction
- citekey-constrained writing bundles
- Word export through Pandoc,
zotero.lua, andtemplate.docx
The server supports four workflow modes through the run_reference_workflow tool:
retrieve_onlyimport_onlyexport_onlyfull_pipeline
Safety Rules
- Never fabricate references, citekeys, DOIs, or citation facts.
retrieve_onlymust not draft prose.import_onlymust not run open-ended literature retrieval.export_onlymust not retrieve literature or add references.full_pipelinekeeps the end-to-end path, but still cannot bypass verified references.- Word export must use Zotero live citations,
zotero.lua, andtemplate.docx.
Requirements
- Python 3.11+
- Zotero Desktop running locally
- Better BibTeX for Zotero
- Pandoc on
PATHfor Word export
Install Python dependencies:
python -m pip install -r requirements.txt
Codex MCP Configuration
Example config.toml entry:
[mcp_servers.cite-rag-mcp]
command = "python"
args = ["C:\\path\\to\\cite-rag-mcp\\server.py"]
startup_timeout_sec = 20
Adjust the path for your local clone.
Main Tool
Use run_reference_workflow.
retrieve_only
Inputs:
citation_need_csv_path, orretrieval_request
Behavior:
- Searches and ranks literature
- Verifies DOI metadata
- Checks Zotero duplicates
- Imports missing verified DOI items into the selected Zotero collection when needed
- Returns citekeys
It does not write body text and does not export Word.
import_only
Inputs:
doi_list, ortitle_listplus optionalauthors_list, orverified_references_csv_path
Behavior:
- Confirms whether items already exist in Zotero
- Imports DOI-backed items when needed
- Returns citekeys or a clear unresolved status
It does not draft body text and does not export Word.
export_only
Inputs:
markdown_path, ormarkdown_content
Behavior:
- Reads finalized Markdown
- Removes any
<Citation_Reasoning>block before rendering - Ensures the references anchor is present
- Calls Pandoc with
zotero.luaandtemplate.docx
It does not retrieve literature and does not add references.
full_pipeline
Inputs:
retrieval_requestorcitation_need_csv_path- Optional finalized
markdown_pathormarkdown_content
Behavior:
- Runs retrieval and verification
- Produces verified citekeys and a citekey-constrained drafting bundle
- Optionally exports Word if finalized Markdown is provided
The actual drafting step should still be performed by the orchestrator or a writing skill using the returned citekey bundle.
Notes
- The included
references/custom_journal_catalog.xlsxis used for journal filtering and ranking. - The Zotero local connector is expected at
http://127.0.0.1:23119. - No Zotero credentials or API secrets are required by this server.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.