Deploy every idea. Pay for the winners.
One resource pool for all your apps. VPS economics, per-app isolation, and none of the server management.
The shift
More apps. Not more traffic.
Hosting today assumes a few important services that scale up. AI-assisted building looks different: dozens of experiments, internal tools, client prototypes — and maybe one that matters.
Most platforms optimize for the winner. Lacho is optimized for everything before it.
Build freely. Keep what works. Leave the infrastructure to us.
Product
The best parts of a VPS and containers
Pack many applications into one pool of resources without making them share one process or turning yourself into a server operator.
Your apps share the capacity you pay for. Quiet experiments leave room for the apps that matter.
Each app has its own boundaries, so packing projects together doesn’t mean letting them interfere with each other.
SQLite, Postgres, or Valkey — pick what fits the app. Same addon model for each, no separate servers.
Addons
Without the service sprawl
Typical platforms run your app here, your database there, and connect everything over the network. Lacho adds a database or cache to each app instead — you choose which one.
SQLite
Lightweight and small on memory. SQL database for your app.
Postgres
Full SQL database. Heavier on memory than SQLite.
Valkey
In-memory key-value store for caching and state. Redis-compatible.
JavaScript frameworks
Tested on Lacho today
We only call a framework supported after it builds, runs inside the Lacho runtime, and serves real requests in our end-to-end tests.
Lacho implements the Node APIs these frameworks need; it is not a complete Node.js replacement yet. Next.js Pages and App Router pass our end-to-end tests — not every Next feature is covered yet.
Early access
What works today
The core runtime is working end to end on a single node. We’re opening it carefully while we harden the public product around it.
Alpha access
Free while we learn
Paid plans are disabled during alpha. We haven’t defined the right limits yet, and we won’t invent prices before measuring real compute, storage, and support costs.
$0 today
Try the runtime, deploy experiments, and help shape the limits that eventually become paid plans.
Idle apps sleep instead of holding active memory. The runtime is designed for wake in under 100 ms, but alpha access has no latency SLA. Future plans will give you a pool of capacity to share across many apps—not charge each experiment as a separate server.
Blog
Notes from the runtime
Why this model of hosting exists, and what we’re building next.
Postgres and Valkey join SQLite as dashboard addons, Next.js passes our end-to-end tests, and wake time is bounded by the slowest part of the app—not the sum of every service.
AI changed how apps are built. Hosting hasn’t caught up.Modern infrastructure prepares every idea for millions of users. AI-assisted building needs the opposite—less complexity until an idea proves itself.