Early Access

A cloud for your agent.

Build and run web apps, websites, agents, databases and background workers on Uniac, the cloud platform for AI agents.

Make your agent build with Uniac

Agents read uniac.ai/agents.md

Then build anything
Works with
Cursor Claude Code GitHub Copilot Codex Cline Windsurf Google Antigravity Kiro Trae Qoder
The platform

From an idea to an app
you can run.

Build with your agent, deploy with Uniac, and stay connected to what's running.

01 Build

Describe your app.
Bring every piece together.

Your agent defines the application and the infrastructure it needs in one file: APIs, background workers, databases, queues and storage.

Reusable

Define a service once and instantiate it under any name. Every deployment in the project ships in one run.

From one prompt for a searchable document library, Claude defines the API, worker, Postgres database, Redis queue, and search index, then runs uniac deploy. All five services are running with private networking at the example URL https://document-library.example.com.

02 Deploy

Ship the next version.
See every step.

Follow each deployment from image preparation to running instances. See task progress, service status, and the version that's running in the platform.

The project graph shows the API sending jobs through Redis to the worker. Both services connect to Postgres and the search index. The API's detail window opens over the graph, showing deployment v2 progressing through preparation, instance startup, verification, and activation. The completed deployment is Active with one running instance out of one requested, a succeeded task, and the Public HTTP endpoint https://document-library.example.com.

03 Operate

See what's running.
Know what it's using.

Follow deployments, inspect your services, and track resource usage and spending from one place.

Stay close to your application.
Then build what's next.

The document-library project contains five running services. The platform shows the worker's resource metrics, including 0.24 vCPU and 128 MiB of memory in this example, and its current deployment, version 02. The example's monthly usage totals $2.84 across five services: $1.24 CPU, $0.86 memory, $0.60 volumes, and $0.14 egress.

Pricing

Pay for what you use.

Your services meter what they actually consume, drawn from credit you already hold.

Subscription
Trial
$2 credit
At sign-up · good for 30 days
Monthly
$5 / month
$5 of usage included every month
Usage-based
CPU $0.000008 / vCPU-second
Memory $0.000004 / GiB-second
Volumes $0.25 / GiB-month
Egress $0.15 / GiB

Metered while your services run — compute by the second, storage by the hour — and drawn from your credit.

Start deploying →
Included
  • No per-project or per-service fees
  • Private networking between services
  • Web dashboard for managing services
  • Community support

Shipping at scale? [email protected]

Watch your idea unfold.

Start deploying →