Citecite Mcp

io.github.MukundaKattaPublished by GitHub account MukundaKattav0.1.1Updated Oct 24, 2026

Inject, parse, and strip [N] citation markers in RAG outputs.

VerifiedSTDIODesktop onlyKnowledge & MemoryAI & ML

Overview

AI-generated overview

Lets an assistant add, read, or remove [N] citation markers in RAG-generated text.

What you can do
A small local MCP server for working with bracketed citation markers in retrieval-augmented generation output. Its tools inject [1] [2] [3] markers at the end of a text or at a given offset, parse every [N] marker in source order with position and length, and strip all markers from a text. It is a text post-processing helper rather than a retrieval or search system.
When to use it
Use it when a RAG pipeline has produced an answer and you need to attach source markers to it, extract the markers already present for post-processing, or clean them out before showing the text. It suits agents that handle citation formatting as a discrete step.
Requirements
Runs locally as a stdio process, started with npx from the npm package @mukundakatta/citecite-mcp, so Node.js and network access for the package fetch are needed. No accounts, API keys, environment variables, or headers are declared. Desktop MCP clients only.
Before you install
The tools only transform text passed to them; no credentials, payments, or third-party data sharing are declared. Text sent to the tools is processed by the local process, so avoid passing content you would not run through a local helper.

Installation

In SourceWeft

  1. Open Citecite Mcp in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

@mukundakatta/citecite-mcp

MCP server for RAG citation markers. Inject [N] markers into LLM output, parse them back when post-processing, or strip them entirely.

Tools

  • inject_citations(text, citations, at?) — append [1] [2] [3] markers at the end of text, or insert at a position offset.
  • parse_citations(text) — return every [N] found in source order with position, idx, and marker length.
  • strip_citations(text) — remove every [N] marker.

Install — Claude Desktop

~/Library/Application Support/Claude/claude_desktop_config.json:

json
{  "mcpServers": {    "citecite": {      "command": "npx",      "args": ["-y", "@mukundakatta/citecite-mcp"]    }  }}

Install — Cursor / Cline / Windsurf / Zed

Same shape in each tool's MCP config. Transport is stdio over npx -y @mukundakatta/citecite-mcp.

Use case

You ran a RAG pipeline. The LLM produced an answer. The retrieval layer told you which source chunks were used. Now the LLM (a second pass, or the same agent) needs to inject [1] [2] markers tied to those sources, or to parse markers back out of an existing piece of text. That's what this is for.

License

MIT OR Apache-2.0.

Source: packages/citecite-mcp/README.md at commit 6fad5bb

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v0.1.1LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts