The database your team can carry. One .qkey file contains your schema, data, and the rules QKeyDB uses to interpret them, so the same database behaves the same on a laptop, a server, or a machine that's never touched the internet.
Meet QKey FormatUse .qkey when humans need to author, diff, and approve schemas and contracts. Switch to compact .qkb when scans and throughput matter—binary storage stays compatibility-aware so upgrades stay predictable.
Bundle separate databases in one portable container. Domains stay isolated, deployment stays simple — no scattered files, clear boundaries.
Declarative @schema, @version, and @fields keep contracts next to the data — readable in PRs, consistent on laptops, servers, or air-gapped hosts.
Move data through JSON, CSV, XML, readable QKey, and binary QKB using the same import, export, and convert workflows—no separate “integration SKU.”
Lightweight text/binary parsing and stringify paths suit browsers and glue services—ideal when partners only need dependable encode/decode at the edge.
Packaged deployments add streaming conversion, comment-preserving round trips, custom format hooks, and richer typed values—built for heavier pipelines without abandoning the same mental model.
// conversational queries
find listings where status is active
show tables
// transactions in plain English
start transaction
add into users name "Alice" email "alice@co.com"
commit transaction
// or drop into SQL any time
execute sql SELECT * FROM orders WHERE total > 500SQL, natural language, the SDK, CLI, and HTTP API all converge at the same canonical planner and executor. No behavioural drift between interfaces — what works in the REPL works in production code, every time.
The NLQ engine scores each parsed query for confidence, surfaces auto-suggestions when intent is ambiguous, and caches query shapes with an LRU cache — so repeated patterns get faster without you touching a thing.
const db = await QKeyDB.connect('app_db', {
file: './data/app.qkey'
});
const rows = await db.table('listings')
.find({ status: 'active' })
.orderBy('price', 'asc')
.limit(10);All three SDKs expose the same capability surface — query execution, ops visibility, LivingIndex controls, governance status, and process monitoring. Pick your stack; the feature set does not change.
Enable strict mode at connect time to block unbounded table scans and large-offset pagination before they reach production. Call .explain() on any query to get the execution plan, index gaps, warnings, and recommended actions.
Set up access with keys and roles from day one — not bolted on after the fact. Move from open development wiring through full audit coverage into enforced deny-by-default policies.
An in-memory LivingIndex keeps reads fast with append-oriented, checkpoint-aware durability. Add secondary indexes when queries repeat — visibility tools flag replay pressure before it bites.
Vector search lives inside the same single-file model — add it when your queries outgrow traditional indexes. It's there when you need it.
Streams ongoing changes fact-by-fact with hash validation, backpressure signals, and RPO ≤ 30 seconds. Built for live replica topologies where lag is observable and stale reads are not acceptable.
Packages snapshots on a schedule — hourly, daily, or weekly — with a deterministic apply path and a staged restore cycle: pause, apply, verify, resume, with abort safety at every step.
Use RellSync for continuous HA, LiteSync for off-site cold copies, or run both together. No external message broker, no sidecar processes — replication lives inside the same artifact model.
Use precision SQL where it matters and natural-language helpers where accessibility wins—every surface routes through the same predictable artifact model.
Lifecycle commands stay inside the file discipline—checkpoint-aware startup and governance presets keep regulated teams oriented without bespoke tooling.
Cluster editions keep topology, trust, routing, and rebalance planning in the cluster file — datasets stay portable. Scale up without rearchitecting what's already working.
Portable artifacts, proven interchange, ACL-first security, indexing that grows with you—and a cluster story when multi-node reality hits.