Deploy in your environment.
Use a repeatable workflow.

Astrolift AI provides a control plane for deploying and operating applications and agents. Declare workload requirements in astrolift.toml and use the CLI or operator interface. AWS is supported today; other backends are in progress.

AWSSupported cloud path today

ManifestDeclarative workload configuration

MITSelf-hostable core

SupportImplementation and operation scoped separately

astro — deploy

astro deploy --app payments# reading astrolift.toml … building image …→ pushing to your registry            ok→ applying to cluster prod-eu (your AWS)  ok→ preview env per PR ready            ok# live in your account — nothing left ours.

Capabilities

Deployment and operations for your workloads

Use shared configuration and operating workflows across applications.

Declarative workloads

Describe resources and routes in astrolift.toml. Backend support and migration requirements depend on the target environment.

Application services

Configure databases, queues and caches through the supported deployment integrations.

Preview environments

Configure isolated environments for pull requests and their cleanup policy.

Tenant and project boundaries

Manage tenants, projects and workloads through the control plane.

Questions about running your own cloud?

Talk to the team — we’ll help you get Astrolift shipping on your own infrastructure.

hello@calliope.ai Visit calliope.ai

What's inside

Open-source components

Astrolift is open source under calliopeai on GitHub. Read it, run it, fork it.

Control plane

The service that owns tenants, projects, workloads and policy — backed by Postgres & Temporal.

Django · GraphQL

calliopeai/astrolift-app View repo ↗

Operator UI

The console for operators — deployments, workflows, secrets, forms and audit in one place.

Next.js

calliopeai/astrolift-app View repo ↗

astro CLI

A single Go binary to authenticate and manage workloads from your terminal or CI.

calliopeai/astrolift-cli View repo ↗

Status page

A lightweight health-monitoring service for your deployments and the control plane itself.

FastAPI

calliopeai/astrolift-status View repo ↗

How it works

Configure and deploy

Set up the control plane, declare the workload and verify it in the target environment.

Install on your cluster

Review the supported backend and provision the control plane with the required database, credentials and networking.

Describe your apps

Declare workload resources and routes in astrolift.toml.

Deploy your way

Use the astro CLI, operator interface or API for your configured deployment workflow.

◇ astrolift.toml

# one manifest — any cloud
[app]
name = "payments"
cloud = "aws"      # or azure | gcp | oci

[service.api]
build = "./api"
port  = 8080
scale = { min = 2, max = 10 }

[resources]
postgres = "15"
redis    = "7"

[previews]
per_pull_request = true

Get started

Install the `astro` CLI.

Pick your platform. Then astro login and ship.

HomebrewmacOS · Linux

brew install calliopeai/tap/astro

ScoopWindows

scoop install astro

curlany shell

curl -fsSL https://astrolift.ai/install.sh | sh

Dockercontainer

docker run --rm calliopeai/astro:latest

Read the docs Star on GitHub ↗

Deployment support

AWS today, additional backends in progress

Workloads run in your selected environment. Network egress depends on application connections, model endpoints and your network policy. Scope on-premises or isolated deployments with the team.

Part of the Calliope AI platform

Build it. Run it. Govern it.

Astrolift is the run layer of Calliope AI — the same team, the same private-AI perimeter, three composable products.

Build

Calliope AI Workbench

Code, notebooks, chat & agents in one governed environment.

Run

Ship the whole stack into your own cloud — multi-cloud, multi-tenant PaaS.

You are here ▲

Govern

Zentinelle

Agent GRC — observability and real-time policy for everything you run.

Deploy in your environment.Use a repeatable workflow.