01 / WEB APPS & APIS

Deploy web apps and APIs.
Stop paying for sleeping staging environments.

Push a Dockerfile or OCI image, pick your cores and region (us-east, eu-central, us-west, ap-south), and get instant HTTPS, private service DNS, and automatic sleep-to-zero on idle.

shell — deploy web app with auto-sleep
HTTPS + AUTO-SLEEP
# Deploy an API service in us-east with instant HTTPS + scale-to-zero
$ uncage workloads create \
    --project proj_prod \
    --name billing-api \
    --surface app \
    --image ghcr.io/acme/billing-api:v2.14 \
    --vcpu 2 --mem 4096 --port 8080

✔ billing-api running   (2 vCPU · 4 GB · us-east)
  Public URL:   https://billing-api.app.uncage.sh
  Private DNS:  billing-api.proj-prod.internal:8080
  Active Rate:  $0.0936/hr   |   Sleeping Rate: $0.0000/hr
Regions: us-east · eu-central · us-west · ap-south Open in Console →
Cold Wake Latency
< 250ms
Idle Compute Cost
$0.00/hr
TLS Provisioning
Automatic
Regions
4 Global
CAPABILITIES

Built for production APIs and ephemeral PR previews.

APP-01 Zero YAML

Dockerfile & Auto-Detect Builds

Deploy directly from your Dockerfile, Git repository, or existing OCI registry image. Node, Go, Python, Rust, and Ruby projects build with layer caching out of the box.

APP-02 Auto TLS 1.3

Instant HTTPS & Custom Domains

Every exposed service receives an encrypted HTTPS endpoint immediately with automatic certificate renewal, HTTP/2, WebSocket pass-through, and custom domain CNAME support.

APP-03 Zero-config DNS

Private Project Networking

Apps, PostgreSQL databases, Redis caches, and worker services inside the same project communicate over private internal DNS without exposing database ports to the public internet.

APP-04 $0.00/hr idle

Scale-to-Zero Preview Environments

Spin up a full isolated stack per pull request. Environments automatically sleep after your configured idle timeout and wake transparently on the next incoming HTTP request.

APP-05 $0.036/vCPU-hr

Predictable Active-Usage Metering

No surprise per-seat platform surcharges or bloated tier jumps. Dial in 1 to 8+ vCPUs and pay strictly per active compute second plus $0.10/GB-month for attached NVMe volumes.

APP-06 One Workspace

Unified With Databases & Sandboxes

Co-locate your production API next to dedicated PostgreSQL 17 and AI agent sandboxes in the exact same region and private network—no cross-vendor latency or separate bills.

REAL-WORLD WORKFLOWS

Where teams cut infrastructure spend and deployment friction.

PR PREVIEWS

Full-stack preview URLs for every pull request

The Pain

Leaving 15 staging and preview containers running 24/7 burns $300+/mo on idle RAM and CPU nobody touches after 6 PM.

With uncage.sh

Set idle_timeout_seconds=300 on preview apps and pair each one with an instant copy-on-write database branch.

→ 92% lower staging bill; previews wake on the first click in under 250ms.
MICROSERVICES

Burst-ready APIs, webhooks, and background workers

The Pain

Traditional PaaS tiers force you to pay for peak memory all month just to absorb periodic webhook or report-generation spikes.

With uncage.sh

Allocate 4 vCPU · 8 GB for snappy request execution while keeping standby workers asleep at $0/hr between batches.

→ Sub-second throughput under load without paying for 24/7 over-provisioning.
AI BACKENDS

Co-located SaaS backends + agent sandboxes

The Pain

Running your web app on one PaaS, Postgres on a second vendor, and AI sandboxes on a third adds latency, IAM glue, and three invoices.

With uncage.sh

Deploy your Next.js or Go API, managed Postgres, and sandbox fleet in one uncage.sh project sharing private DNS.

→ Single API key, unified logs, and sub-millisecond internal service calls.
TRANSPARENT APP ECONOMICS

What real web workloads actually cost on uncage.sh.

Every web app uses the exact same active-usage meter: $0.036/active vCPU-hr, $0.0054/active GB-hr, $0.10/GB-mo NVMe disk, and $0.00/hr while sleeping.

Open Interactive Pricing Calculator →
PR Preview Environment
1 vCPU · 1 GB · 10 GB disk
1.5 hrs/day active, sleeps otherwise
Estimated $2.86 / mo
Internal Admin / Staging API
2 vCPU · 4 GB · 20 GB disk
8 hrs/day weekdays (176 hrs/mo)
Estimated $18.47 / mo
Always-On Production API
2 vCPU · 4 GB · 20 GB disk
24/7 active (730 hrs/mo)
Estimated $70.33 / mo
App Compute Spec NVMe Disk Active Rate Sleeping Rate Deploy
1 vCPU · 1 GB 10 GB NVMe $0.0414/hr $0.0000/hr Launch →
2 vCPU · 4 GB 20 GB NVMe $0.0936/hr $0.0000/hr Launch →
4 vCPU · 8 GB 40 GB NVMe $0.1872/hr $0.0000/hr Launch →
8 vCPU · 16 GB 80 GB NVMe $0.3744/hr $0.0000/hr Launch →
WEB APPS FAQ

Common questions about deploying apps on uncage.sh.

01 How does scale-to-zero work for incoming HTTPS traffic?
When an app has an idle timeout configured and receives no requests for that window, its memory and disk state are paused and compute billing drops to $0.00/hr. When the next HTTPS request arrives at your endpoint, the edge router holds the connection briefly, wakes the workload in under 250ms, and proxies the request seamlessly.
02 Can I keep production APIs always-on with zero cold starts?
Yes. Set idle_timeout_seconds to 0 for always-on production services. You still benefit from low active-usage rates ($0.036/vCPU-hr + $0.0054/GB-hr) with zero per-seat fees.
03 How do web apps talk to managed databases securely?
Every workload in a project is automatically registered on private internal DNS (for example, pg-prod.my-project.internal:5432). Traffic between your app and database stays on your isolated project network and never traverses the public internet.
04 Do you support WebSockets, SSE streaming, and long-running LLM responses?
Yes. The edge HTTPS proxy natively supports HTTP/1.1, HTTP/2, WebSockets, and Server-Sent Events (SSE) with no arbitrary 10-second serverless timeout cuts.
SHIP IN 60 SECONDS

Deploy your first web app with zero idle tax.

Pick your cores and region, push your image, and stop paying for sleeping dev and preview stacks.

✓ $10 starter credit on signup
✓ Strictly $0.00/hr compute when sleeping
✓ Sub-second boot across 4 global regions