Choose a workflow

Choose the interface that matches your data and the result your application needs. You can combine graph, vector, and key-value operations around a shared entity identifier.

I want to…	Use	Start here
Store and retrieve documents	Key-value records containing bytes, JSON, or another application format.	Multimodal walkthrough
Find related entities	Cypher patterns over nodes and relationships.	Cypher
Find similar text or images	Client-generated embeddings in a named vector index.	Vector search
Combine similarity, relationships, and payloads	A shared entity UID across vector, graph, and KV data.	Multimodal and hybrid queries
Query triples	SPARQL over loaded subjects, predicates, and objects.	SPARQL
Track moving objects	SpaceKey observations with position, timestamp, and entity ID.	Geospatial and spatiotemporal indexing
Import tabular or ontology files	Dataset inspection, reviewed loading, and optional ontology materialization.	Import a dataset
Query external Parquet data	Reviewed lakehouse mappings scoped to a database.	Register, review, and query Parquet
Ask domain questions in natural language	A semantic profile plus an MCP client or semantic console.	Natural-language questions
Control access	Multi-tenant identity, roles, and policies.	User Administration
Diagnose slow or failed requests	Health, query plans, traces, and namespace reports.	Observability
Understand a selected plan	Parsing, semantic eligibility, physical selection, snapshot binding, and runtime work.	Query planning and execution

I want to…

Use

Start here

Store and retrieve documents

Key-value records containing bytes, JSON, or another application format.

Multimodal walkthrough

Find related entities

Cypher patterns over nodes and relationships.

Cypher

Find similar text or images

Client-generated embeddings in a named vector index.

Vector search

Combine similarity, relationships, and payloads

A shared entity UID across vector, graph, and KV data.

Multimodal and hybrid queries

Query triples

SPARQL over loaded subjects, predicates, and objects.

SPARQL

Track moving objects

SpaceKey observations with position, timestamp, and entity ID.

Geospatial and spatiotemporal indexing

Import tabular or ontology files

Dataset inspection, reviewed loading, and optional ontology materialization.

Import a dataset

Query external Parquet data

Reviewed lakehouse mappings scoped to a database.

Ask domain questions in natural language

A semantic profile plus an MCP client or semantic console.

Natural-language questions

Control access

Multi-tenant identity, roles, and policies.

User Administration

Diagnose slow or failed requests

Health, query plans, traces, and namespace reports.

Observability

Understand a selected plan

Parsing, semantic eligibility, physical selection, snapshot binding, and runtime work.

Query planning and execution

Choose storage before loading

Import records when the application will update them through ShrikeDB. Register a lakehouse source when Parquet files remain the authoritative data. Use SpaceKey for append-only geographic observations; graph property writes do not populate a SpaceKey index automatically.

For shared deployments, choose the authentication mode before onboarding users and data. See feature availability for configuration requirements and HA deployment status for clustered workloads.