NEW

SurrealDB 4 is here, introducing SurrealDB Earth and SurrealDS!

Learn more

SurrealDB vs.
Postgres

Migrate from PostgresContact us
PostgreSQL. Relational at its core, with extensions for JSON, full-text and vectors, and no native graph model.
SurrealDB. Relational, graph, vector, document and temporal data natively, with edges traversed without joins.
Speaks Postgres. SurrealDB 4 talks the Postgres wire protocol, so psql, Postgres drivers and ORMs connect as they are.
  1. Babcock
  2. ING
  3. British Airways
  4. Nvidia
  5. Apple
  6. SpaceX
  7. Samsung
  8. adidas
  9. Tencent
  10. Alibaba
  11. PolyAI
  12. Later
  13. Verizon
  14. Liberty Mutual
  15. Walmart
  16. Carrier
  17. Saks Fifth Avenue
  18. San Francisco Compute Company
  19. Shield AI
  20. Wix

01 |THE DIFFERENCE

Where Postgres reaches, SurrealDB runs natively

Postgres covers much of this through extensions and joins. SurrealDB carries every model in the core, in one transaction.

COMPATIBILITY

Your Postgres clients, connected

SurrealDB 4 speaks the Postgres wire protocol, so psql, Postgres drivers and ORMs such as Prisma, Drizzle and TypeORM connect without a new client, and most queries run unchanged.

Read the protocol docs
RELATIONSHIPS

Traversal in place of joins

Postgres reaches a related row through a foreign key and a join, and the cost grows with depth. SurrealDB follows the edge directly, with no join planning and no join index.

Explore graph relations
VECTORS

Embeddings beside the rows

pgvector adds a vector column and an index type to a relational core. SurrealDB stores the embedding with the record and queries it alongside graph, full-text and structured filters in one statement.

Explore vector search
SCALE

Horizontal reads and writes

Postgres scales past one node through primary and replica coordination, and large deployments shard by hand. SurrealDB separates compute from storage and scales both directions without sharding.

Explore the architecture

02 |THE SAME QUERY

Similar documents, from one team, this quarter. Three joins and an operator in Postgres with pgvector. One statement in SurrealQL.

Postgres + pgvector

1-- pgvector: the similarity, then three joins
2SELECT d.title,
3       1 - (d.embedding <=> $1) AS score
4FROM   document d
5JOIN   authored a ON a.document_id = d.id
6JOIN   person p   ON p.id = a.person_id
7JOIN   membership m ON m.person_id = p.id
8WHERE  m.team_id = $2
9  AND  d.published > now() - interval '90 days'
10ORDER  BY d.embedding <=> $1
11LIMIT  10;

SurrealQL

1-- SurrealQL: the similarity and the path, together
2SELECT
3    title,
4    vector::similarity::cosine(embedding, $q) AS score
5FROM document
6WHERE embedding <|10|> $q
7    AND <-authored<-person->member_of->team CONTAINS $team
8    AND published > time::now() - 90d
9ORDER BY score DESC;

03 |HOW IT COMPARES

SurrealDB vs. Postgres + pgvector

AI and retrieval workloads stretch a relational engine past its design. SurrealDB handles them natively.

Data and relationships

SurrealDB

PostgreSQL

Models
NativeNative multi-model: document, relational, graph, key-value, time-series, vector, full-text search, and geospatial.
Via extensionsRelational database with JSON, full-text search, and vector support via extensions. No native graph model.
Relationship querying
TraversalFirst-class relationships traversed directly without joins. No join planning or join indexes required.
JoinsRelationships expressed through foreign keys and joins. Cost increases with join depth.
Operations
Architecture
DistributedDistributed, multi-model database with a decoupled query and storage layer. Efficient as single node or distributed.
MonolithicMonolithic relational engine. Extensions bolt additional capabilities onto a core not designed for multi-model execution.
Scale
HorizontalHorizontally scalable for both reads and writes. Designed to eliminate the need for sharding.
ShardedScaling beyond a single node requires manual primary/replica coordination. Large deployments commonly require manual sharding.
Schema evolution
FlexibleSchema-flexible by design. Start schemaless and incrementally enforce schema without downtime.
MigrationsSchema changes on large tables are disruptive and risky. Migrations often require maintenance windows.
Cost
Pricing
UsageOne engine in place of several specialised systems. On SurrealDB Earth a database scales to zero when idle, so you pay for the data it stores.
InfrastructureInfrastructure costs grow with index count, replicas, and sharding.

TRUSTED BY

Enterprise teams building on SurrealDB

See more case studies

Three data stores collapsed into one

Samsung

An AI assistant for 10,000 technicians

Verizon

Nine backend tools consolidated into one

Tencent

Low-latency RAG, under the customer's control

PolyAI

45 million recommendation queries a month

Saks Fifth Avenue

Postgres,answered

Yes. SurrealDB 4 speaks the Postgres wire protocol for a defined subset of SQL, so psql, Postgres drivers and ORMs such as Prisma, Drizzle and TypeORM connect without a new client, and most queries run unchanged. Some column types carry caveats, and defining a schema through an ORM is more limited for now.

THE PLATFORM

Everything an application and its agents know. Five surfaces, one engine.

Database

Document, graph, vector, time-series and relational in one engine.

Five data models as dotted tiles: documents, graph, vector, time-series and relational
Read more

Agent Memory

What an agent learns, with its source and its time, in the same engine.

A timeline of remembered facts, each with its source
Read more

Cloud

Dedicated clusters by region, or serverless on SurrealDB Earth.

Clusters in three regions on a world map, each running or scaling
Read more

Studio

Query, explore and design the schema from the browser.

A SurrealQL query in Studio and the schema graph under it
Read more

MCP

Every model that speaks MCP reaches the database and the memory directly.

Three models connected through MCP to the database and Agent Memory
Read more

IN PRODUCTION

Trusted at scale. Samsung, Nvidia, Verizon, Tencent, and Walmart run on SurrealDB.

14,000+

Developers building on SurrealDB Cloud

4M+

Developers building on SurrealDB worldwide

FROM THE TEAMS

SurrealDB gives us a foundation where we can unify semantic search, knowledge graphs, and AI-driven decision making without stitching together multiple systems. Collapsing responsibility into SurrealDB has become our default engineering posture.
Justin Foley

VP of Engineering, Later

GET STARTED

Keep your Postgres client. Add graphs, vectors and memory in one query, on SurrealDB 4.

Start free with SurrealDBMigration guide