Natural Language → Cypher / SPARQL Translator#

Convert natural language questions into executable graph queries.

This skill translates human-readable requests into graph query languages such as Cypher (Neo4j) or SPARQL (RDF), allowing developers to interact with graph databases using natural language instead of writing complex query syntax.

Quick Start#

Use Cases#

Use this skill when a user wants to:

• ✅ Convert natural language questions into graph queries
• ✅ Generate Cypher queries for property graph databases
• ✅ Generate SPARQL queries for RDF triple stores
• ✅ Explore graph datasets using plain English
• ✅ Quickly prototype graph queries

Example requests:

• "Find employees who work at Acme."
• "Show all companies located in California."
• "List people connected to Alice within two hops."
• "Find products purchased by customers in London."

What This Skill Produces#

The skill generates executable graph queries in supported query languages:

• 🔗 Cypher queries (Neo4j)
• 🔗 SPARQL queries (RDF triple stores)
• 📝 Query explanations
• 🎯 Query parameter templates

Translation Process#

The translator follows a 5-step process to convert natural language into executable queries:

1. Identify Entities — Detect entity names (people, organizations, locations)
2. Identify Relationships — Extract relationship verbs (works at, located in, purchased, etc.)
3. Determine Query Type — Classify request type (nodes, relationships, paths, counts, aggregations)
4. Construct Query Pattern — Build graph patterns representing relationships
5. Generate Query Syntax — Convert patterns into Cypher or SPARQL

For detailed information about each step, see Query Translation Strategy.

Supported Query Languages#

Cypher#

Used in Neo4j and other property graph databases.

Example:

cypher

MATCH (p:Person)-[:PURCHASED]->(product:Product)
RETURN p, product

📖 See Cypher Query Guide

SPARQL#

Used in RDF triple stores and knowledge graphs.

Example:

sparql

SELECT ?person ?product
WHERE {
  ?person :purchased ?product .
}

📖 See SPARQL Query Guide

Common Patterns#

Simple Entity Lookup#

text

"Find all people named Alice"

Relationship Traversal#

text

"Show companies where employees work"

Multi-Hop Queries#

text

"Find people connected to Alice within two relationships"

For more patterns, see Query Patterns Reference.

Examples#

Browse practical examples organized by complexity and use case:

• 🚀 Basic Translations
• 🔀 Multi-Hop Queries
• 🎯 Parameterized Queries
• 📊 Domain-Specific Examples

Best Practices#

When generating and executing graph queries:

✓ Use schema labels consistently

✓ Avoid overly broad relationship patterns

✓ Prefer parameterized queries for security

✓ Include LIMIT clauses for large datasets

✓ Validate query syntax before execution

See Architecture & Implementation for advanced considerations.

Advanced Topics#

• 🏗️ System Architecture & Design
• 🔧 Entity Recognition Pipeline
• 🔗 Relationship Extraction
• 📚 API Reference
• ⚠️ Known Limitations & Edge Cases

Related Skills#

• Graph Query Debugging Tool
• Graph Query Optimization Assistant
• Graph Template Query Generator
• Multi-Hop Reasoning Query Builder

---

Last Updated: 2026-03-08

Maintainer: OpenClaw Community

License: See LICENSE in repository root