GraphQL Editor & Templates
The Query Editor is a Monaco-powered GraphQL interface with intelligent, schema-aware templates that adapt to your collection's actual schema.
Features
- Monaco Editor with full GraphQL syntax highlighting
- Schema-aware autocompletion based on your connected Weaviate instance
- 12+ query templates covering all major Weaviate operations
- Real-time validation and error highlighting
- Dynamic template generation that auto-populates properties from your schema
Query Templates
Choose the right template for your use case:
| Template | Use Case |
|---|---|
| Basic Get | Simple object retrieval |
| Vector Search (nearVector) | Similarity search with explicit vector |
| Semantic Search (nearText) | Concept-driven search with AI embeddings |
| Object Search (nearObject) | Find similar to existing object |
| Hybrid Search | BM25 + Vector with alpha balancing |
| BM25 Search | Keyword-based retrieval |
| Generative Search | RAG with AI-generated summaries |
| Filter Query | Complex WHERE conditions |
| Aggregation Query | Statistics (count, min, max, mean) |
| Group By Query | Aggregate by property categories |
| Relationship Query | Explore cross-references |
| Explore Query | Metadata, vectors, and per-object generation |
| Tenant Query | Multi-tenant data access |
Schema-Aware Generation
When your schema is available, dynamic templates:
- Auto-populate primitive properties (text, number, boolean, date)
- Include geo-coordinates with latitude/longitude selection
- Keep cross-reference selections minimal (inline fragment with
_additional.id) - Tailor generative prompts to available text properties
- Detect embedding model dimensions for 15+ popular models
Configuration
Customize template generation via QueryConfig:
| Field | Purpose |
|---|---|
limit | Result limit (default: 10) |
offset | Pagination offset |
includeVectors | Include object vectors |
includeMetadata | Include creation/update timestamps |
includeScores | Include distance/certainty/score |
maxProperties | Cap displayed properties |
tenantName | Multi-tenant target |
searchQuery | BM25/hybrid query text |
concepts | nearText concepts |
alpha | Hybrid balance (0=vector, 1=keyword) |
Distance vs. Certainty
- distance (preferred in v1.14+): Lower = closer. Threshold acts as maximum.
- certainty (legacy): 0–1 range; higher is stricter.
Supported Embedding Models
The template system detects dimensions for 15+ models:
- OpenAI (ada-002: 1536, text-embedding-3-small: 1536, text-embedding-3-large: 3072)
- Cohere v3: 1024
- Sentence Transformers (mpnet): 768
- BERT, PaLM, Ollama, AWS Bedrock, and more
TIP
For a comprehensive guide covering all templates, best practices, and a troubleshooting FAQ with 41 common mistakes, see the full GraphQL Templates Guide.