Spin up apps for people.

The backend that
gets you to the
App Store.

Auth, payments, push, in-app purchase, real-time, agents — and the store builds to get there. One SDK, one substrate.

Include the SDK  ·  Get weird!
A DJ throwing a hand up for the bass drop, rendered with an RGB-split glitch
AUTHPAYMENTSREAL-TIMEAGENTSOPERATOR PLANEANALYTICSON-DEVICEMANUAL OR AUTOONE SDK
The problem

You shipped an MVP this weekend. Now you need everything else.

Agent-assisted coding changed what one person can build. It did not change what one person can operate.

The moment you need real users, you’re back to stitching a half-dozen services that don’t know about each other — writing glue instead of product.

  • Firebase for auth, but only the basics
  • Stripe for payments, but not in-app purchase
  • Pusher / Ably for real-time
  • Twilio for notifications
  • SendGrid for email
  • Cloudinary for media
  • A LangChain wrapper for agents (good luck)
  • Custom admin tools you wrote at 2 AM
  • No way to pause a misbehaving customer
  • A migration story when you outgrow it
The platform

One substrate. Every primitive your app needs.

BaaS got you a database and auth and called it a backend. We give you the whole operating system.

Auth, pluggable

Turn on OTP, magic link, passkey or OAuth per app — each app gets only what it needs. Standards-based, Better Auth underneath.

Payments, 5 rails

Take money five ways — Apple IAP, Google IAP, Stripe Connect, Paddle, grants — into one ledger that reconciles. Native, not bolt-on.

Real-time, serverless

Live updates with no cluster to babysit and no idle servers in the bill. API Gateway + Lambda + DynamoDB underneath.

Agents, governed

The automatic mode of your app — every LLM call cost-tracked, rate-limited, evaluated and prompt-versioned, so you can see what it spent.

Operator plane

Pause an app. Maintenance mode is a button — not a 2 AM script.

Media, analytics, search

Upload, process and serve media, then ask questions of it in under a second and search the same data. S3 + DuckDB over parquet underneath.

On-device agents

The same agent runtime on the phone as on the server — one framework, no port. Server side ships today; on-device inference (MediaPipe) is in progress.

Scales with you

Handles your growth without rewrites. Same architecture on day one and day one thousand.

Manual or auto

Every app you ship has two modes.

Your users can drive it by hand — or hand the wheel to an agent. Same app, same data, same capabilities; the agent just takes the controls you already built.

That’s not a chatbot bolted in the corner. It’s the automatic transmission for your product — running on the exact substrate the manual app does.

Everyone’s selling “AI-first.” That’s autopilot with no airplane. Build the app first — then build the agent into it.

Manual — the app

The user holds the loop: they work the controls, see the result, decide what’s next. The product you ship on day one.

Auto — the agent

The user states a goal and hands off the loop. The agent pulls the same levers and checks in at the gates. One config flip from the app you have.

The manifesto

Omniscience, the K-maxxer.

One argument, end to end — from what a solution is, to why you now write intent instead of code. Same argument, three altitudes — read it in your language.

↑ pick a register to read it

Pricing

Start free. Turn dials when you scale.

Your plan buys capabilities. Usage is metered separately — build one app or fifty, you pay for what you run, not how you split it.

Free
$0
forever
Build it
  • Unlimited apps — we never charge per app
  • Auth, data, real-time, media, push
  • Store-ready builds — TestFlight, App Store, Play
  • Bring your own AI key
  • We pause you at the limits
  • Community support
Most popular
Indie
$12
per month
Plug in
  • Managed AI — no API key, no provider account
  • Spend caps you set, we enforce
  • Everything in Free, with room
  • Email support
Starter
$49
per month
Get paid
  • Payments, 5 rails — Stripe, Apple IAP, Google IAP, Paddle, grants
  • One ledger, real reconciliation — native, not bolt-on
  • Managed agents, metered at cost
  • Full real-time
Growth
$199
per month
Run it like a business
  • Full operator plane — pause, resume, maintenance mode
  • Agent governance — eval, prompt versioning, cost dashboards
  • Larger included inference credit
  • Analytics + search · priority support
Scale
$599+
per month
Isolate & comply
  • Dedicated tenancy → dedicated cell
  • BYO credentials, Tenant-scoped
  • Custom limits, all metered
  • SLA + dedicated support
Usage · metered

Your plan is capabilities.
Usage is separate.

The unbounded resources — the ones that actually cost us — are metered at provider cost + margin, behind a ceiling you set. Set a cap; we enforce it. No surprise bills.

AI inference
Provider cost + margin. BYO key for zero markup.
Egress & bandwidth
Metered above your plan baseline.
Media processing
Per output-minute of transcode.
Storage & API calls
Baseline included; overage metered per unit.

Prices, included credits, and metered rates are indicative while we finalize projections.

Stop stitching services.
Ship the whole product.

The batteries-included backend for mobile apps. Opinionated on purpose.