WorldCypher Overview

Canonical surface for the Query Engine layer.

ArcFlow's query language is GQL — the ISO international standard for graph query languages (ISO/IEC 39075). If you know Cypher, you already know it: MATCH, CREATE, MERGE, WHERE, RETURN all work exactly as expected, with no migration needed.

ArcFlow's implementation is called WorldCypher. It passes 100% of the openCypher TCK, implements ISO/IEC 39075:2024 (GQL Edition 1) plus accepted forward-looking V2 proposals, and extends GQL with capabilities the standard leaves to implementations.

WorldCypher extensions beyond the standard: AS OF temporal snapshots, LIVE MATCH standing queries, CREATE LIVE VIEW, confidence filtering (WHERE r._confidence > 0.8), spatial predicates, CREATE SKILL, PROCESS NODE.

WorldCypher is a language for world models. A single query can express all three dimensions of agent reasoning simultaneously: spatial (what's near me, what's in this zone), temporal (what was true at sequence N, what changed since), and relational (what connects these entities, through what paths, with what confidence). That composability is what separates world model queries from SQL or key-value lookups.

Operator Summary#

Quick Examples#

Read#

MATCH (n:Person) WHERE n.age > 25 AND n.name CONTAINS 'li' RETURN n.name

Write#

CREATE (n:Person {name: 'Eve', age: 28}) RETURN n
MERGE (n:Person {name: 'Eve'}) RETURN n
MATCH (n) WHERE n.name = 'Eve' SET n.age = 29
MATCH (n) WHERE n.name = 'Eve' DELETE n

Traverse#

MATCH (a:Person {name: 'Alice'})-[:KNOWS*1..3]->(b) RETURN b.name
shortestPath (a:Person {name: 'Alice'}), (b:Person {name: 'Dave'})

Aggregate#

MATCH (n:Person) RETURN count(n), avg(n.age), collect(n.name) ORDER BY n.age DESC

Introspect#

EXPLAIN MATCH (n:Person) WHERE n.age > 25 RETURN n.name

Sections#

Query Syntax#

• MATCH — pattern matching
• WHERE — filtering predicates
• RETURN — projection, ordering, pagination

Statements#

• CREATE — node and relationship creation
• MERGE — upsert semantics
• DELETE — removal
• SET — property mutation

Data Types#

• Overview — property value types

Functions#

• Aggregations — COUNT, SUM, AVG, COLLECT
• shortestPath — graph algorithms
• EXPLAIN — query plan introspection

Execution#

• Execution Options — QueryOptions(deadline_ms=…), result.transport_outcome, result.io_stats

Standards & Compatibility#

• GQL Conformance — standards lineage, TCK results, comparison with other implementations
• GQL & WorldCypher Reference — complete syntax and feature matrix