chiron-common-core-standards

An MCP (Model Context Protocol) server that exposes US Common Core ELA standards (grades 2–4) as lookup tools for LLM agents.

Package name on PyPI: chiron-common-core-standards. Python import name: chiron_standards.

Backed by a bundled SQLite database with full-text and semantic search — no network calls at runtime, no external rate limits.

Why

When an LLM generates educational content — worksheets, assessments, lesson plans — it needs to align to real educational standards. Without a reliable lookup tool, models hallucinate standard codes, misremember descriptions, or attach the wrong grade level. chiron-standards gives agents a grounded, deterministic source for Common Core lookups.

Features

• Three tools:
  • get_standard(code_or_id) — exact lookup by notation (e.g. "RL.3.1") or GUID
  • list_standards(jurisdiction, grade, subject) — full tree for a standard set
  • search_standards(query, grade, subject, limit) — hybrid FTS5 + semantic search with Reciprocal Rank Fusion
• Offline — all ~500 Common Core ELA standards (grades 2–4) ship inside the wheel as a pre-built SQLite file with embeddings
• Semantic search powered by OpenAI text-embedding-3-small (only search_standards requires an OPENAI_API_KEY; the other two tools are fully offline)
• Leaf-only search results — parents, domains, and anchor standards are filtered out so agents align to alignable standards

Install

pip install chiron-common-core-standards

Use with Claude Desktop

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

{
  "mcpServers": {
    "chiron-standards": {
      "command": "chiron-standards",
      "env": {
        "OPENAI_API_KEY": "sk-..."
      }
    }
  }
}

Restart Claude Desktop. The three tools will appear automatically.

Use with the MCP Inspector (for development)

uv run fastmcp dev inspector src/chiron_standards/server.py

Open the printed URL, connect, and call tools manually.

Use as a Python library

The same tools are available as plain Python functions — useful from your own backend without going through MCP:

from chiron_standards.db import connect
from chiron_standards.queries import get_standard, search_standards

conn = connect()

# Exact lookup
print(get_standard(conn, "RL.3.1"))

# Semantic search (requires OPENAI_API_KEY in env)
hits = search_standards(conn, "finding evidence in a passage", grade="03", subject="ELA", limit=3)
for h in hits:
    print(h["notation_short"], h["description"])

Data model

Each standard returns a dict with:

search_standards results additionally include a _score field (higher = better).

Build the database yourself

The published wheel already contains standards.db. To rebuild it (e.g. to pull fresh upstream data):

git clone https://github.com/JovannyEspinal/chiron-common-core-standards
cd chiron-common-core-standards
uv sync
cp .env.example .env   # add CSP_API_KEY and OPENAI_API_KEY
uv run python -m chiron_standards.ingest

This fetches Common Core ELA grades 2–4 from the Common Standards Project API, embeds descriptions with OpenAI, and writes src/chiron_standards/data/standards.db.

Development

uv sync
uv run pytest                           # run the test suite
uv run fastmcp dev inspector src/chiron_standards/server.py

License and attribution

• Code: Apache License 2.0 — see LICENSE.
• Data: The standards data shipped in this package is sourced from the Common Standards Project, which aggregates releases from the Achievement Standards Network (ASN) operated by Desire2Learn. ASN's standards data is licensed under CC BY 3.0 US.
• Common Core State Standards © 2010 National Governors Association Center for Best Practices and Council of Chief State School Officers. All rights reserved.

See NOTICE for full attribution requirements.