A new MCP server for AI agents, alongside an updated public API and Python library, make FOLIO's legal ontology easier than ever for courts and legal organizations to adopt.
We are excited to announce a new MCP server for the FOLIO (Federated Open Legal Information Ontology) project, alongside significant updates to the free public REST API and Python library — all designed to make it easier for legal organizations of any size to build better information systems.
Try It Now
The public API is live at folio.openlegalstandard.org with interactive documentation at /docs. No API key required.
”Garbage in, garbage out” has always been true, but the consequences are sharper in legal systems where the quality of information directly affects access to justice. Courts, legal aid organizations, and small firms all deal with the same fundamental challenges: classifying documents, categorizing areas of law, identifying entities, and connecting related concepts.
Larger organizations can afford to build custom taxonomies and invest in knowledge management. Smaller ones — the municipal courts, solo practitioners, legal aid clinics, and community organizations that serve millions of people — typically cannot. The result is fragmented, inconsistent data that makes it harder to find relevant information, share knowledge across organizations, or build reliable AI systems.
FOLIO is an open knowledge graph with over 18,000 concepts covering areas of law, document types, legal entities, governmental bodies, courts, and more. It is licensed under CC-BY 4.0, meaning anyone can use it freely.
The latest round of updates — a new MCP server and improvements to the existing API and Python library — make it even more practical to use FOLIO in real systems.
The folio-mcp package is a new MCP (Model Context Protocol) server that makes FOLIO available as a set of tools for AI assistants. It works with Claude Code, Claude Desktop, Cursor, and any MCP-compatible client.
Adding FOLIO to Claude Code takes one command:
claude mcp add folio -- uvx folio-mcp
This gives AI agents the ability to search, browse, and retrieve ontology concepts during conversations — useful for tasks like classifying documents, tagging case types, or validating legal metadata.
The MCP server includes 12 tools, 3 resources, and prompt templates for common classification tasks. It can run in API mode (instant startup, uses the public API) or local mode (loads the full ontology in-process for offline use).
The FOLIO API is a free REST API with open CORS origins, so it can be called from any web application, backend service, or script. There is no API key and no rate limiting for normal use. Recent updates have added multi-provider LLM search, an interactive explorer UI, and improved search result highlighting.
The API supports:
For organizations that want to self-host, the folio-api repository includes Docker support and a built-in explorer UI.
The folio-python library provides programmatic access to the full ontology. Recent updates have added structured query filtering, multi-provider LLM search with effort/tier configuration, and improved search indexing including language-tagged labels:
pip install folio-python
from folio import FOLIO
folio = FOLIO()
# Search by prefix
results = folio.search_by_prefix("Mich")
for owl_class in results:
print(f"Class: {owl_class.label}")
# Get all areas of law
for area in folio.get_areas_of_law():
print(area.label)
The library supports fuzzy matching, prefix search, structured queries with composable filters, and LLM-powered semantic search across multiple providers (via the [search] extra). It can also export concepts in OWL XML, JSON-LD, or Markdown.
We built these tools with a specific audience in mind: the courts, legal aid organizations, small firms, and public interest groups that need better data but cannot afford to build custom knowledge management systems.
A few concrete examples:
All FOLIO resources are free, open source, and available now:
We welcome contributions, feedback, and questions. If your organization is interested in adopting FOLIO or building on these tools, please reach out through the FOLIO website or open an issue on any of the GitHub repositories.