Lux Cloud is a managed application database powered by Lux: one Rust engine for tables, cache, vectors, auth, realtime, queues, and time series.
Build the whole backend around application state that can live, move, search, react, expire, queue, and stream without splitting the product across half a dozen services.
Tables, cache, vectors, auth, realtime, queues, and metrics managed together from a single product surface.
Create a project, copy your SDK keys, run migrations, watch logs, and ship without assembling infra first.
The things modern apps need are not bolt-ons. Lux runs them in one engine so they can compose.
Powered by the MIT-licensed Lux runtime: Rust, Redis-compatible, WAL-backed, snapshot-capable, and self-hostable.
The stack keeps getting wider
Redis Cache
Pinecone Vectors
Kafka Realtime
BullMQ Queues
Lux CloudModern products are not just rows. They are users, sessions, live state, semantic memory, background jobs, metrics, queues, and cache. Lux Cloud gives those primitives one managed home.
Lux Cloud sells the managed product. The Lux engine is the moat: every primitive runs close to the data, speaks through one runtime, and becomes available from one SDK.
table().select()
Typed app data with indexes, joins, foreign keys, migrations, and a familiar fluent SDK.
SET / GET / TTL
Redis-compatible hot state for sessions, rate limits, counters, feature flags, and ephemeral data.
KSUB / Pub/Sub
Listen to key changes and channels without running a separate realtime service.
VSET / VSEARCH
Semantic retrieval and agent memory live beside the operational records they reference.
XADD / XREADGROUP
Jobs and workflows built into the same runtime as the product state that triggers them.
TSADD / TSRANGE
Metrics and temporal events with retention, aggregation, labels, and direct query access.
/v1/project
Browser, serverless, and edge access through Cloud gateway keys and user sessions.
lux.auth
App users, sessions, JWTs, OAuth providers, and project-aware access from the same SDK.
The Lux thesis
The old stack treats cache, vectors, queues, realtime, auth, and metrics as different products. Lux treats them as different shapes of application state.
That means fewer moving parts, but more importantly, it means the pieces can start to compose like one database.
Realtime app data
Subscribe to the rows, keys, and channels that drive your UI instead of polling or wiring a second service.
AI-native records
Store a message, attach an embedding, search nearby vectors, and render matching records from one project.
Auth-aware products
Create users, issue sessions, call the Cloud gateway with project keys, and build app tables around identity.
Operational workflows
Let writes, queues, metrics, and realtime events live close enough to feel like one system.
Use the project SDK for Cloud apps, Redis-compatible clients for direct protocol access, HTTP for edge/serverless calls, and the CLI for migrations and operations.
import { createBrowserClient } from "@luxdb/sdk"
const lux = createBrowserClient(
"https://api.luxdb.dev/v1/my-project",
"lux_pub_..."
)
// auth and user sessions
await lux.auth.signInWithOAuth({ provider: "github" })
// tables
await lux.table("messages").insert({ channel_id: "general", body: "hello" })
const { data, error } = await lux.table("messages").select().eq("channel_id", "general").limit(50)
// vectors and time series live in the same project
await lux.vectorSet("doc:1", embedding, { row_id: 1 })
const matches = await lux.vectorSearch({ vector: queryVec, k: 5 })
await lux.tsAdd("latency:p95", 42.8, { labels: { route: "/api/messages" } })
Works with
Node.jsPythonGoRustRubyJava.NETBullMQPricing
$10 / month / project. The whole application database.
Tables, auth, cache, vectors, realtime, queues, time series, storage, snapshots, logs, and dashboard. Free for 14 days, no credit card required.
Powered by the open-source Lux engine · redis-benchmark · 50 clients · 1M requests · pipeline 64
SET
3.4x faster
GET
2.6x faster
LPUSH
3.9x faster
PUBLISH
6.6x faster
"Lux Cloud is built around a simple idea: the things your app needs most should not be scattered across six products. They should feel like one database."
Why the engine matters
Read more →Auth, tables, cache, vectors, realtime, queues, metrics, and durable storage belong in the same backend instead of being stitched together service by service.
