In-memory or persistent? Yes. Our new in-memory storage engine. Lock-free.
Reads never write to shared memory, so they keep scaling as you add cores. 21.51 million point reads per second across 15 million keys, on a 32-core, 64-thread AMD Ryzen Threadripper 9970X with 128 GB of RAM and an NVMe SSD. It runs embedded in an application, in the browser and on a server, and can persist to disk when the workload requires it.
- 28.3×
- Faster than RocksDB in-memory for reads
- 17.5×
- Faster than LMDB in-memory for reads
- 6.61×
- Faster than LMDB in-memory for updates
Blog
Benchmark
Whitepaper
Documentation
GitHub
Speed or durability? Yes. Our new persistent storage engine. We are replacing RocksDB.
It will be the only persistent storage engine in SurrealDB 4, fully durable and fully transactional. 6.80 million point reads per second across 500,000 keys with every commit flushed to disk, on a 32-core, 64-thread AMD Ryzen Threadripper 9970X with 128 GB of RAM and an NVMe SSD.
- 7.95×
- Faster than RocksDB for reads, fully durable
- 26.8×
- Faster than RocksDB for creates, fully durable
- 73.7×
- Faster than SlateDB for reads, fully durable
Blog
Benchmark
Whitepaper
Documentation
GitHub
Memory or accuracy? Yes. SurrealDB Agent Memory. Memory you can finally trust.
Every fact keeps its history, its permissions and its source, in one SurrealDB engine. Know what changed, when it changed, who can see it and where it came from.
- 77.2%
- BEAM 1M, Gemini 3.5 Flash Lite judge, 700 questions
- 57.1%
- BEAM 1M questions with every rubric item fully met
Blog
Benchmark
Deep‑dive
Documentation
Get started
Stable or new? Yes. SurrealDB 4. Out now.
A rebuilt query core, PostgreSQL and MySQL compatibility, and more than 240 fixes that no 3.x release contains. Every 3.x record stays readable in SurrealDB 4.
- 240+
- Fixes no 3.x release contains
- 90+
- Community bug reports 3.x did not fix
- 4,500+
- SurrealQL test files, run on every change
SurrealDB 4
Documentation
GitHub
Get started
Simplicity or scale? Yes. SurrealDB Earth. Elastic from zero, serverless on Cloudflare.
Serverless SurrealDB on Cloudflare's network, fully transactional. Each database is its own instance, with its own storage and commit log. It starts with the first query, cold starts in ~300ms and scales to zero instantly when idle, with its data stored natively on object storage. It scales to as many agents and users as you need, wherever they are, and you pay for requests and storage, with a free tier that needs no card.
- 10+ billion
- Peak rows per second on SurrealDB Earth. Transactional.
SurrealDS. Every write stored in two zones.
SurrealDS is distributed storage for SurrealDB on object storage, in SurrealDB Cloud multi-node clusters and Enterprise Edition. With the default two-zone policy, a write is acknowledged only after it is stored on disks in both zones, with object storage such as Amazon S3 holding the durable copy. If a node fails, a replacement node recovers from the surviving copy and object storage. Query nodes keep caches while data pages live in object storage, so a node joins without copying the dataset.
SQL or SurrealQL? Yes. SurrealDB 4 speaks PostgreSQL and MySQL.
PostgreSQL and MySQL clients connect over their own wire protocols and run a defined subset of PostgreSQL and MySQL SQL, through the same query planner as SurrealQL. Most queries run unchanged, and anything outside the supported set fails with a named error. Use SQL where it fits and SurrealQL where it goes further, on the same data and the same permissions, and run code written for the MongoDB Node.js driver through mql.js.
Transactions or speed? Yes. SurrealDB 4. A rebuilt query core, still ACID.
A rebuilt query core with faster indexing, faster graph traversal and faster vector search, and every query still fully ACID.
- 7.4×
- Faster vector index builds
- 1.1 ms
- Filtered vector search, down from 27 ms
- 0.19 ms
- Counting a 100,000-edge vertex, down from 37 ms
Stop choosing.
Build with all of it.
Start for free In the cloud
A free SurrealDB Cloud instance, ready in minutes.
SurrealDB Earth Serverless on Cloudflare
Isolated instances that start on their first query.
Install SurrealDB On your machine
Run SurrealDB locally with one command.
Read the docs Start building
Guides and reference for SurrealDB and SurrealQL.
Agent Memory Start in Studio
Memory for agents, on the same engine as the database.
Talk to us Plan your deployment
Work through your architecture with the SurrealDB team.
THE PLATFORM
Everything an application and its agents know. Five surfaces, one engine.
Database
Document, graph, vector, time-series and relational in one engine.

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

Cloud
Dedicated clusters by region, or serverless on SurrealDB Earth.

Studio
Query, explore and design the schema from the browser.

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

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.
VP of Engineering, Later
