Home

The AI-Native
Bitemporal Database

Append-only immutable storage with MVCC snapshot reads, full SQL, LSM-tree engine, and embedded AI runtime. Built in Rust.

276ns Point Reads

1.9µs Point Writes

4× Faster reads vs SQLite

20× Faster writes vs SQLite

Get Started SQL Reference

Why TensorDB?

Every write is an immutable fact. Every fact has a timeline. No data is ever lost, overwritten, or silently mutated.

Full SQL Surface

DDL, DML, JOINs, CTEs, window functions, aggregates, 60+ built-in functions. Real SQL — not a query DSL.

Learn more →

Bitemporal by Default

Separate system time (when recorded) from business time (when valid). Time-travel with AS OF and VALID AT.

Learn more →

Vector Search

Native vector similarity search with cosine, euclidean, and dot-product distance metrics. Build RAG pipelines in-database.

Learn more →

Sub-Microsecond Performance

276ns point reads (4× faster than SQLite). 1.9µs point writes (20× faster). Direct shard bypass, lock-free fast path.

Learn more →

LSM Storage Engine

WAL with CRC framing, skip-list memtable, prefix-compressed SSTables, bloom filters, LRU caches, multi-level compaction.

Learn more →

Performance

TensorDB sled redb SQLite

Point Read Latency (lower is better)

TensorDB

276ns

sled

244ns

redb

573ns

SQLite

1.1µs

Point Write Latency (lower is better)

TensorDB

1.9µs

sled

9.2µs

redb

18.4µs

SQLite

38.6µs

Prefix Scan (100 keys, lower is better)

TensorDB

45µs

sled

38µs

redb

89µs

SQLite

156µs

Mixed Workload (80% read, 20% write)

TensorDB

2.1µs

sled

4.8µs

redb

7.2µs

SQLite

12.4µs

Benchmarks run on Linux x86_64, single-threaded, 1M pre-loaded keys. Lower is better for all measurements. See full benchmarks.

Try TensorDB SQL

TensorDB SQL

SELECT — Basic QueryINSERT — Immutable FactAS OF — Time TravelVALID AT — Bitemporal QueryJOIN — Cross-Table QueryEXPLAIN — Query Plan

SELECT name, email, created_at
FROM users
WHERE status = 'active'
ORDER BY created_at DESC
LIMIT 5;

Results 5 rows

name	email	created_at
Alice Chen	alice@example.com	2025-03-15 09:22:01
Bob Smith	bob@example.com	2025-03-14 14:05:33
Carol Wu	carol@example.com	2025-03-14 08:47:19
Dave Park	dave@example.com	2025-03-13 22:11:45
Eve Jones	eve@example.com	2025-03-12 16:33:08

INSERT INTO transactions (id, account, amount, currency)
VALUES ('txn_001', 'acc_alice', 250.00, 'USD');

-- Every write becomes an immutable fact with:
--   commit_ts  = auto-assigned system timestamp
--   valid_from = now (or specified)
--   valid_to   = infinity

Results 1 row

result
1 row inserted (commit_ts: 1042)

-- Read the database as it was at commit 500
SELECT account, balance
FROM accounts
AS OF 500
WHERE account = 'acc_alice';

Results 1 row

account	balance
acc_alice	1,500.00

-- What was the policy rate valid on 2024-06-15?
SELECT policy_id, rate, valid_from, valid_to
FROM insurance_policies
VALID AT 1718409600
WHERE policy_id = 'POL-2024-001';

Results 1 row

policy_id	rate	valid_from	valid_to
POL-2024-001	4.25	1704067200	1735689600

SELECT u.name, COUNT(o.id) as order_count,
       SUM(o.total) as total_spent
FROM users u
JOIN orders o ON u.id = o.user_id
GROUP BY u.name
HAVING total_spent > 100
ORDER BY total_spent DESC;

Results 3 rows

name	order_count	total_spent
Alice Chen	12	2,450.00
Bob Smith	8	1,823.50
Carol Wu	5	967.25

EXPLAIN ANALYZE
SELECT * FROM orders
WHERE user_id = 'u_alice'
  AND status = 'shipped';

Results 7 rows

plan
Filter { predicate: status = 'shipped' }
└─ PointLookup { key: 'u_alice', cost: 1.00 }
execution_time_us: 42
rows_returned: 3
bloom_filter_hits: 1
cache_hits: 2

Get Started

cargo add tensordb

Read the quickstart →View on GitHub

License: TensorDB is available under the PolyForm Noncommercial License 1.0.0. Free for personal, educational, and research use. Contact us for commercial licensing.