Query and explore
Run read-only graph queries, inspect paths in the browser, and export the results when you need the full dataset.
Run against a published model
Open Query (查询), select a published model, and choose Cypher, SQL, or Gremlin. In the Parquet demo, this Cypher query summarizes two-hop transfers to each destination.
MATCH (a:Account)-[:Transfer]->(b:Account)-[t:Transfer]->(c:Account)
WHERE a.tenant_id = 1 AND a.account_id = 10
RETURN c.account_id AS account, count(*) AS paths, sum(t.amount) AS amountUse the query options to set a timeout or provide Cypher parameters. The result identifies the model revision and source versions. Check these when reproducing a result.
Investigate the graph
Return a path or graph objects to populate the graph view. Select a node or edge to inspect its properties. Expand a node to see its neighbors and load additional batches as needed. Use the table view for scalar results.
MATCH p = (:Account)-[:Transfer]->(:Account)
RETURN pWork with larger results
The workbench shows a bounded preview. Use cancellation for a query you no longer need. For complete results, enable full-result capture and download Arrow IPC with query and export permissions. Use filters, hop bounds, and resource limits to keep exploratory queries predictable.
Applications can use HTTP, Bolt, Gremlin WebSocket, the CLI, or embedded Rust. See the query and protocol reference and the graph algorithms reference for shortest paths, PageRank, and community detection.