Vector 3.0 — branching databases are here
Ship realtime features before lunch.
Vector is the sync engine, database and edge runtime in one — so every screen of your product updates the instant your data does.
Trusted by product teams at
The platform
Everything realtime. Nothing to wire up.
Realtime by default
Every query is a live subscription. When a row changes, every screen that shows it updates — no sockets, no cache invalidation, no polling.
Edge-fast reads
Reads are served from the region nearest your user, with a p95 under 40ms worldwide.
Branch your data
Copy-on-write branches of production in under two seconds, one for every pull request.
Typed end to end
Your schema generates the client. Rename a column and your editor knows before your users do.
Global in 35 regions
Multi-region by default, with automatic failover and data residency you can choose per table.
Secure from row one
Row-level policies written in TypeScript, audited on every change and enforced at the edge.
Developer experience
From empty folder to production in four commands.
A local sync engine that behaves exactly like production, branches that copy nothing, and one deploy that reaches every region. Your terminal is the whole control panel.
- npm create vector@latest ✔ Project acme created — schema.ts, vector.config.ts
- vector dev ● Sync engine on localhost:4000 · 3 live queries
- vector branch create checkout-v2 ✔ Branched production in 1.8s · 2.4M rows, zero copies
- vector deploy --prod ✔ Live in 35 regions · p95 38ms
How it works
One model. Every screen live.
Model your data
Describe tables, relations and policies in one TypeScript file. Vector migrates, types and documents it for you.
Subscribe from any screen
Swap a fetch for a live query. The component re-renders when the data changes — anywhere in the world.
Branch for every pull request
Each preview gets its own copy of production data, created in seconds and deleted on merge.
Ship to the edge
One command deploys functions and data rules to 35 regions, with instant rollbacks.
We deleted eleven thousand lines of cache code in the first month. Our dashboards finally agree with each other.
Branch-per-pull-request changed how we review. Designers click a link and see real data, not fixtures.
The only infrastructure change our customers noticed was that everything got faster.
Pricing
Start free. Scale when it’s working.
Hobby
$0per month
- 1 project
- 500k synced events
- Community support
- Shared regions
- Preview branches
No card needed.
Pro
$29per seat / month
- Unlimited projects
- 50M synced events
- Branch per pull request
- Choose your regions
- Email support in 4 hours
Cancel any time.
Scale
Customannual agreement
- Dedicated clusters
- Data residency controls
- SSO & audit logs
- 99.99% uptime SLA
- A named engineer
Volume pricing available.
Questions, answered.
Does Vector replace my database?
It can. Vector is a Postgres-compatible database with sync built in — or it can sit in front of the Postgres you already run and stream its changes.
How does pricing work when we grow?
Pro includes 50 million synced events per month, then $1 per additional million. Scale plans bundle volume at a lower rate.
Can we self-host?
Scale customers can run Vector in their own cloud account, managed by us, in any of the three major providers.
Which frameworks are supported?
React, Vue, Svelte, Swift and Kotlin have first-party clients. Anything that speaks HTTP can use the REST and streaming APIs.
Is my data used to train anything?
Never. Your data is yours, encrypted at rest and in transit, and deleted within 30 days of closing an account.
Book a demo
See your own data go live in thirty minutes.
Bring a schema or a hard problem. An engineer — not a salesperson — will walk you through Vector on a call and leave you with a working branch.