Skip to main content
MCP server CX Codex

Docket Lens — Claude Code MCP server for FDA docket analysis

Docket Lens is an MCP server that retrieves FDA docket comments and attachment text, then compares draft and final guidance to trace submitter requests and document changes.

MIT updated 2026-10-28

Open on GitHub

What Docket Lens does

Docket Lens searches FDA dockets and returns metadata, documents, comments, and extracted attachment text. Results are kept in a local SQLite cache; the server reaches the network on a cache miss or when refresh=true. You can gather comments into a resumable corpus and export it as JSONL.

For guidance comparisons, the server splits documents into sections, matches the draft against the final version, and marks paragraph edits and moves. It uses no model calls for this work. The comparison measures alignment, not impact, so it cannot establish that a comment caused a change. Setup requires Python 3.11 or later and a Regulations.gov API key; DEMO_KEY is too rate-limited for a full corpus pull.

The repository also includes analysis scripts and a web explorer with linked views of submitters, requests, and document changes. The server uses the MIT license; the explorer has a separate PolyForm Noncommercial 1.0.0 license.

What it helps with

  • search FDA dockets by keyword
  • retrieve docket comments and extracted attachment text
  • build a resumable comment corpus
  • compare draft and final guidance section by section
  • export docket data as JSONL for skills and subagents

Example prompts for the agent

  • Find FDA dockets about risk assessment and show the comment count.
  • Build a corpus for FDA-2020-D-2307, including attachments.
  • Compare the draft and final guidance for this docket and show the changed sections.

How to install Docket Lens

Install

The MCP registry has no launch command for it: see the repository README

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about Docket Lens

How do I install Docket Lens in Claude Code?

Install Python 3.11 or later, then run git clone https://github.com/Black-Swan-Causal-Labs/regintel-mcp.git, cd regintel-mcp, python3 -m venv .venv && .venv/bin/pip install -e ., and cp .env.example .env. Put your Regulations.gov key in .env as REGULATIONS_GOV_API_KEY=..., then register the server from the repository directory with claude mcp add regintel-mcp -- "$(pwd)/.venv/bin/python" -m regintel_mcp.

What does Docket Lens do?

It retrieves FDA docket comments and their attachments, then compares draft guidance with the final text. The comparison shows alignment and document changes, not whether a comment caused a change.

Do I need an API key to build a full corpus?

Yes. Add a Regulations.gov key to .env as REGULATIONS_GOV_API_KEY=.... DEMO_KEY is too rate-limited for a full corpus pull.

Similar MCP servers

MCP servers catalog →

Project sponsors

Companies that keep this analytics open