LegalMCP
LegalMCP is a comprehensive US legal MCP server that enables AI assistants to search over 4 million US court opinions, manage Clio practice data, and access PACER federal filings through natural language.
README
LegalMCP
The first comprehensive US legal MCP server for AI assistants.
Connect Claude, GPT, Cursor, or any MCP-compatible AI to 4M+ US court opinions, Clio practice management, and PACER federal filings. Research in seconds, not hours.
30-Second Demo
You: Find Supreme Court cases about Fourth Amendment and cell phone location data
LegalMCP: Found 52 results. Top case:
Carpenter v. United States, 585 U.S. 296 (2018)
The Court held that accessing historical cell-site location
information constitutes a search under the Fourth Amendment,
requiring a warrant supported by probable cause.
→ 127 cases cite this opinion
→ Full text: courtlistener.com/opinion/4578834
Why LegalMCP?
| Traditional Research | AI + LegalMCP | |
|---|---|---|
| Find relevant cases | 45-90 min | < 30 sec |
| Trace citation history | Open Westlaw, click around | "Who cited this case?" |
| Pull client billing | Log into Clio, navigate menus | "Total hours on Henderson?" |
| Monthly cost | $200-400 (Westlaw/Lexis) | Free |
Quick Start
Install
We recommend using a virtual environment to avoid conflicts with other packages:
# Create and activate a virtual environment
python -m venv legal-mcp-env
# Windows
legal-mcp-env\Scripts\activate
# Mac/Linux
source legal-mcp-env/bin/activate
# Install
pip install legal-mcp
Or install from GitHub:
pip install git+https://github.com/Mahender22/legal-mcp.git
Run
Want to try it without API keys? Enable demo mode first (optional):
# Mac/Linux
export LEGAL_MCP_DEMO=true
# Windows
set LEGAL_MCP_DEMO=true
Then start the server:
legal-mcp
Connect to Claude Desktop
Add to your claude_desktop_config.json:
Windows (%APPDATA%\Claude\claude_desktop_config.json):
{
"mcpServers": {
"legal-mcp": {
"command": "C:/path/to/legal-mcp-env/Scripts/legal-mcp.exe",
"env": {
"LEGAL_MCP_DEMO": "true"
}
}
}
}
Mac/Linux (~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"legal-mcp": {
"command": "/path/to/legal-mcp-env/bin/legal-mcp",
"env": {
"LEGAL_MCP_DEMO": "true"
}
}
}
}
Note: Use the full path to
legal-mcpinside your virtual environment. Remove theLEGAL_MCP_DEMOline and add your API keys for real data (see SETUP.md).
Connect to Claude Code
Run this command to add LegalMCP globally (available in every session):
Mac/Linux:
claude mcp add legal-mcp /path/to/legal-mcp-env/bin/legal-mcp
Windows:
claude mcp add legal-mcp C:\path\to\legal-mcp-env\Scripts\legal-mcp.exe
To enable demo mode, add the env flag:
Mac/Linux:
claude mcp add legal-mcp -e LEGAL_MCP_DEMO=true -- /path/to/legal-mcp-env/bin/legal-mcp
Windows:
claude mcp add legal-mcp -e LEGAL_MCP_DEMO=true -- C:\path\to\legal-mcp-env\Scripts\legal-mcp.exe
Tip: To add it to a specific project only, add
-s projectflag or create a.mcp.jsonfile in your project root.
Connect to Cursor / Windsurf
LegalMCP works with any MCP-compatible client. Add the legal-mcp command to your AI tool's MCP server configuration.
Configuration
Set environment variables for API access. See SETUP.md for step-by-step instructions.
| Variable | Required | Description |
|---|---|---|
COURTLISTENER_TOKEN |
Optional | Higher rate limits for case law search |
CLIO_TOKEN |
For Clio tools | OAuth token for practice management |
PACER_USERNAME |
For PACER tools | PACER account username |
PACER_PASSWORD |
For PACER tools | PACER account password |
LEGAL_MCP_DEMO |
Optional | Set true for demo mode (no API keys) |
All 18 Tools
Case Law
| Tool | What It Does |
|---|---|
search_case_law |
Search 4M+ US court opinions by topic, court, date range |
get_case_details |
Get full opinion text for a specific case |
get_case_record |
Get docket — parties, judges, procedural history |
find_citing_cases |
Find cases that cite a specific opinion |
find_cited_cases |
Find cases that an opinion relies on |
parse_legal_citations |
Parse Bluebook citations from any text |
list_available_courts |
List all 400+ courts and their codes |
list_reporter_abbreviations |
Decode reporter abbreviations (U.S., F.3d, etc.) |
Practice Management — Clio
| Tool | What It Does |
|---|---|
search_clients |
Search contacts by name, email, phone |
search_matters |
Search matters by number, description, status |
get_matter_details |
Full matter info — client, billing, deadlines |
get_time_entries |
Billable hours by matter, attorney, date range |
get_matter_tasks |
Tasks and to-dos for a matter |
get_matter_documents |
Documents attached to a matter |
get_calendar |
Hearings, deadlines, and meetings |
Court Filings — PACER
| Tool | What It Does |
|---|---|
search_federal_cases |
Search PACER for federal court cases |
get_federal_case |
Get case details from PACER |
get_court_filings |
Get docket entries and filings |
Note: PACER charges $0.10/page for document downloads. LegalMCP searches and lists for free — it does not auto-download paid documents.
Pricing
Free and open source. All 18 tools, MIT license, no paywall. Just install and go.
Docker
# Copy and configure environment
cp .env.example .env
# Edit .env with your API keys
# Run
docker-compose up
This starts the MCP server on port 8000 and the waitlist API on port 8080.
How It Works
LegalMCP is a Model Context Protocol server. MCP is an open standard that lets AI assistants call external tools — like searching case law or querying your Clio data.
┌──────────────┐ MCP ┌──────────────┐ API ┌──────────────┐
│ │ ──────────► │ │ ──────────► │ │
│ Claude / │ Tool calls │ LegalMCP │ HTTP │ CourtListener│
│ GPT / │ ◄────────── │ Server │ ◄────────── │ Clio / PACER │
│ Cursor │ Results │ (your PC) │ JSON │ │
└──────────────┘ └──────────────┘ └──────────────┘
Your data stays on your machine. LegalMCP runs locally and connects directly to the APIs.
Development
# Clone and install
git clone https://github.com/Mahender22/legal-mcp.git
cd legal-mcp
pip install -e ".[dev,waitlist]"
# Run tests
pytest legal_mcp/tests/ -v
# Run server locally
python -m legal_mcp.src.server
License
MIT — use it however you want.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。