Made for your AI agent.
A planned CLI and API let your agent deploy a project, check status, read logs and return its public URL. No server setup between you and an idea.
A home for the apps you make with AI. Deploy from your agent, get a public URL, and run multiple apps inside one predictable monthly workspace.
Interactive demo · No real deployment or payment
“Deploy my app and give me a URL.”
$ pangea deploy . --name my-app
✓ Build complete
✓ App started
✓ HTTPS route ready
https://my-app.pangea.hostIllustrative CLI and output. The CLI and hostname are not live.

A planned CLI and API let your agent deploy a project, check status, read logs and return its public URL. No server setup between you and an idea.
Keep a side project, a small API and an internal tool together. Allocate CPU and memory per app within a shared workspace budget.
Automatic HTTPS and a Pangea subdomain are part of the intended live experience. Custom domains are planned. Public access is an explicit deployment choice.
YOUR APPS / INTERACTIVE PREVIEW
Deploy, inspect, stop and restart sample apps. Everything here is simulated; no source code, credentials or payment details are requested.
React app · 0.25 vCPU · 256 MiB
idea-board.pangea.hostURL previewNode.js API · 0.25 vCPU · 256 MiB
daily-api.pangea.hostURL previewAPPS WORKSPACE / PROPOSED PRICING
Separate from CI build-minute plans. Pay for a workspace resource pool; no per-app subscription within that pool. Prices are proposals in USD before taxes.
Proposed trial: one small app.
1 vCPU + 2 GiB across your apps
2 vCPU + 4 GiB across your apps
CPU, memory and storage are shared workspace limits, not dedicated physical hardware. Fixed resource limits apply to each running app. New allocations are blocked when the pool is full; data writes fail at the disk limit. Public traffic pauses at the transfer allowance until reset or an explicit plan change; there is no automatic overage charge. Monthly plans renew until canceled, with access through the paid period. Pricing, build limits and live onboarding are still being designed. The demo models CPU and memory allocation only; it does not provision containers, storage, domains or subscriptions.
BUILDS AND APPS, UNDER ONE ROOF
CI jobs finish and release their resources. Apps keep running to serve requests. Pangea is designed to give both a home, with separate limits and clear monthly plans.
Explore CI plansMultiple sample apps, pooled CPU and memory, agent deployment playback, URL previews, logs and stop/start controls.
What comes before live accessAgent authentication, isolated builds and runtimes, public routing, secrets, persistent storage, resource enforcement and real online billing. Private service-to-service networking and managed databases are future options.