Built in Rust. Runs Everywhere.
One database to replace them all
SQL, vectors, documents, graph, and compute -- from a single binary. Zatabase ships with a PostgreSQL wire protocol, a high-performance similarity engine, and a built-in merchant payment rail with KYC. No sidecar services. No glue code.
Try it now -- no signup required. Insert data, run queries, and see semantic search in action.
Request alpha access
Zatabase is invite-only during the alpha. Tell us what you'd build with it.
Stop managing five databases
PostgreSQL for relations, MongoDB for documents, Redis for caching, Pinecone for vectors, Neo4j for graphs. Zatabase consolidates all five into one platform with a single operational surface.
Inside the binary
...and here's how it works
One binary, one network roundtrip, every database surface your stack already speaks. Watch a request flow from your app through the wire protocol into the router, then fan out across the engines.
Clients
Protocols
Gateway
Engines
One engine, many query languages
Write SQL, vector similarity, document queries, or graph traversals -- Zatabase parses and executes them all natively.
CREATE TABLE products (
id UUID PRIMARY KEY,
name TEXT,
price DECIMAL(10,2),
tags JSONB
);
SELECT name, price
FROM products
WHERE price > 29.99
ORDER BY price DESC
LIMIT 10;CREATE TABLE articles (
id UUID PRIMARY KEY,
title TEXT,
embedding VECTOR(384)
);
-- KNN with metadata filtering
SELECT title,
embedding <-> ARRAY[0.1, 0.2, ...] AS dist
FROM articles
ORDER BY dist LIMIT 10;# Stream NDJSON into a collection
curl -X POST /v1/collections/events/ingest \
-H "Content-Type: application/x-ndjson" \
--data-binary @events.ndjson
# Query with projections
curl /v1/collections/events/query \
-d '{"filter": {"type": "purchase"},
"project": ["user_id", "amount"],
"limit": 50}'-- Connect with psql, pgAdmin, DBeaver...
psql "postgresql://user:pass@your-project.zatabase.io/mydb"
-- Full transaction support
BEGIN;
INSERT INTO ledger (account, amount)
VALUES ('acct_1', -100.00);
INSERT INTO ledger (account, amount)
VALUES ('acct_2', 100.00);
COMMIT;Platform
Everything in one binary
Stop stitching together pgvector, Redis, Elasticsearch, environments, backups, and a job queue. Zatabase replaces them all with a single, coherent system.
Semantic search, built in
Find by meaning, not just keywords. ML embeddings, similarity scoring, and approximate-nearest-neighbor indexing -- no separate vector database to run.
Full SQL, full speed
Joins, transactions, window functions, EXPLAIN -- the SQL you already know, with native vector and document operators baked into the same engine.
Schema-flexible documents
Stream JSON straight in, query it back as tables. No upfront migrations, no separate document store. Same database, two shapes.
Drop-in Postgres, no SDK
Whatever already speaks Postgres works with Zatabase. psql, pgAdmin, DBeaver, every language driver, every ORM. Zero rewrites.
Real-time over data channels
Stream queries, change events, and metrics through WebRTC. Sub-millisecond latency for dashboards, multiplayer apps, and live agents.
Cached at the edge, automatically
Hot data stays near your users. Spin up an edge node through Zatabase, earn credits while it serves requests, and pay only for the queries that miss the cache.
Payments & Credits
Built-in metered billing, merchant payment rail, and KYC verification. Stripe integration for self-serve top-ups. One transparent platform fee per transaction.
Dev, staging, prod -- one click
Spin up isolated environments per branch or per release. Promote a snapshot to staging, branch a database from prod for a clean repro -- in seconds.
Point-in-time backups
Continuous incremental backups with point-in-time restore. Roll back a table, a project, or the whole org without paging anyone.
Org-scoped multi-tenancy
Organizations, teams, projects, and RBAC built in. Every request scoped, every byte isolated. Onboard customers without standing up a new database.
Connect from anywhere
Use the PostgreSQL wire protocol, the REST API, or a native SDK. Pick your language and start querying in seconds.
# Connect via PostgreSQL wire protocol
psql "postgresql://user:password@your-project.zatabase.io/mydb"
# Run a query
SELECT name, embedding <-> ARRAY[0.1, 0.2, 0.3] AS dist
FROM products
ORDER BY dist LIMIT 10;Built for demanding workloads
From AI inference pipelines to financial ledgers, Zatabase handles the use cases where generic databases fall short.
AI and Machine Learning
Store embeddings alongside structured metadata in a single table. Run KNN queries that filter on relational predicates and rank by vector distance in one pass -- no ETL pipeline required.
- ✓ RAG retrieval with metadata filtering
- ✓ Real-time embedding ingestion via NDJSON
- ✓ HNSW index tuning per collection
Fintech and Payments
Full ACID transactions with BEGIN/COMMIT/ROLLBACK via the PostgreSQL wire protocol guarantee consistency for ledger operations. A built-in payment rail with metered billing and merchant verification ships in the box.
- ✓ Serializable isolation for double-entry ledgers
- ✓ Stripe-integrated credit top-ups
- ✓ Org-scoped RBAC permissions
IoT and Edge
A single static binary with no external dependencies runs on edge devices and cloud servers alike. Stream telemetry via WebRTC data channels with gzip-compressed NDJSON batches.
- ✓ Single binary, zero runtime dependencies
- ✓ CDC triggers for event-driven pipelines
- ✓ Mesh networking for multi-node clusters
Simple, transparent pricing
Start free. Scale on demand. No hidden fees.
Community
Self-hosted, single node
- ✓ Full SQL + vector engine
- ✓ PG wire protocol
- ✓ Document collections
- ✓ Community support
Pro
Managed cloud, multi-tenant
- ✓ Everything in Community
- ✓ Managed cloud hosting
- ✓ Multi-org support
- ✓ Priority support
Enterprise
Dedicated infrastructure
- ✓ Everything in Pro
- ✓ Dedicated clusters
- ✓ SLA + on-call support
- ✓ SSO + audit logs
Ready to simplify your stack?
Replace five services with one. Zatabase is a cloud-native database platform, written in Rust, that deploys as a single binary. Get started in under two minutes.