RushDB vs Weaviate: graph relationships vs single-hop references.

Side by side

The same task in both: define structure and write connected data.

Both snippets get connected, searchable order data into the database. The difference is how much structure you must declare before the first write.

import weaviate
from weaviate.classes.config import Property, DataType

client = weaviate.connect_to_local()

# Production guidance from Weaviate's own docs:
# disable auto-schema and define collections manually
client.collections.create(
    'Order',
    properties=[
        Property(name='summary', data_type=DataType.TEXT),
        Property(name='status', data_type=DataType.TEXT),
    ],
)

# Plus: configure a vectorizer per collection, and model
# links as single-hop cross-reference properties —
# a bi-directional link needs two reference properties.

from rushdb import RushDB

db = RushDB('RUSHDB_API_KEY')

db.ai.indexes.create({'label': 'ORDER', 'propertyName': 'summary'})

# No collection definition — types are inferred on write,
# and nesting creates traversable relationships
db.records.import_json(
    label='CUSTOMER',
    data={
        'name': 'Acme',
        'ORDER': [
            {'summary': 'Annual plan upgrade', 'status': 'paid'},
            {'summary': 'Added 40 seats', 'status': 'pending'},
        ],
    },
)

Switching

Moving from Weaviate: keep the hybrid-search habits, gain a real graph.

Teams usually hit the wall when retrieval needs connected context — the thing cross-references only approximate one hop at a time. Migration is straightforward because Weaviate objects are JSON: exports import directly as records, and the links you modeled as reference properties become traversable relationships.

Who this is for

Pick the right tool for the job.

If your retrieval needs are mostly single-collection hybrid search, and you want the ecosystem depth of a long-established open-source vector database with 63 official integrations, Weaviate is a strong choice. If your product needs multi-hop relationships as a first-class query primitive (not a denormalization workaround) alongside semantic search, RushDB’s graph-first model fits better.

Comparison table

Data model, search, and agent tooling

Weaviate rows below are sourced from Weaviate’s own docs and pricing page, checked on the date shown. Note that "no multi-hop graph traversal" is an inference from the absence of such documentation, not an explicit statement by Weaviate.

Weaviate facts sourced from weaviate.io, docs.weaviate.io, and Weaviate’s GitHub license file as of the dates shown above; the client snippet follows Weaviate’s documented collections.create shape. The "no multi-hop traversal" claim is an inference from documentation absence, flagged as such. RushDB facts are drawn from RushDB’s own docs and feature pages linked throughout this site. Report a stale claim to hi@rushdb.com.

Implementation example

Store connected records and recall them by meaning

Relationships come from the write path itself — no separate cross-reference property to model.

from rushdb import RushDB

db = RushDB("RUSHDB_API_KEY")

db.ai.indexes.create({
    "label": "MEMORY",
    "propertyName": "output",
})

db.records.create(
    label="MEMORY",
    data={
        "agent_id": "agent-42",
        "kind": "decision",
        "output": "Keep the free plan, but increase Pro onboarding limits after launch feedback.",
        "status": "approved",
    },
)

recall = db.ai.search({
    "labels": ["MEMORY"],
    "propertyName": "output",
    "query": "What did we decide about Pro onboarding?",
    "where": {"agent_id": "agent-42", "status": "approved"},
    "limit": 5,
})

print(recall.data)

Questions developers ask.

Related

Go deeper on the workflow.