KnowJudges

What KnowJudges MCP server does

The KnowJudges MCP server exposes six focused research tools for U.S. judges, courts, cases, courtroom presentation guidance, and recorded rule changes. Results are designed to retain source information, including official decision links when available. The service is intended for AI assistants, research agents, and other software that can use MCP.

Its research scope includes public judge profiles, sourced biographical information, court presentation requirements, cases involving AI-generated legal errors, and changes recorded since KnowJudges began tracking them. The returned material should be treated according to its sources: judge-specific facts should not be presented as courtwide guidance, and sourced information should not be converted into unsupported behavioral claims.

How it works

The KnowJudges MCP server uses Streamable HTTP. Connect an MCP client to https://mcp.knowjudges.com/mcp; the client can then discover the available tools, validate inputs, and request structured KnowJudges results. Public access currently does not require an authentication token.

A typical judge-research sequence starts with find_judges, using information such as a name, court, or jurisdiction to identify the likely person and canonical KnowJudges slug. An agent can then call get_judge_profile and get_judge_biography, followed by courtroom-rule research when relevant. For legal-AI research, find_ai_abuse_cases returns up to 10 cases and may include the judge, court response, dataset source, and a verified official decision link.

Setup and configuration

Use the remote URL as the server address. A client that accepts an MCP server configuration object can use:

{
  "mcpServers": {
    "know-your-judge": {
      "url": "https://mcp.knowjudges.com/mcp"
    }
  }
}

Some clients ask only for a URL; in that case, enter the same endpoint. Configuration field names differ between clients, so follow the client’s MCP setup format.

The public MCP service shares request limits with the public KnowJudges REST API: 20 requests per hour and 100 requests per day for each public client or network. Initialization, tool discovery, and tool calls each count as HTTP requests. Exceeding either limit returns HTTP 429.

Tools and capabilities

• find_judges: Resolve a likely judge and return the canonical KnowJudges slug.
• get_judge_profile: Retrieve a public profile with source-linked, scoped findings.
• get_judge_biography: Retrieve sourced public background information without behavioral inference.
• get_courtroom_presentation_rules: Research requirements and guidance involving PowerPoint, video, exhibits, demonstratives, and related courtroom presentations.
• find_ai_abuse_cases: Find up to 10 court cases involving AI-generated legal errors.
• get_recent_rule_changes: Retrieve changes recorded since KnowJudges change tracking began.

The KnowJudges MCP server is most suitable when the consuming application preserves returned provenance, source URLs, and verified official decision links rather than summarizing beyond the evidence.

Limitations and notes

The public service has shared hourly and daily request limits, which can affect repeated, batch, or recurring research. The available material is source-linked research, not a basis for inferring judicial behavior. Courtwide or jurisdiction-wide guidance must remain distinct from information about an individual judge. Higher-volume usage, recurring feeds, datasets, and private integrations are directed to the KnowJudges MCP documentation.