Elastic infrastructureforAI agents

    Bring your stack. Scale to millions.

    Deploy without rewrites

    Bring Node.js, Python, PHP, and more without rewriting for edge-specific runtimes.

    Sandboxed execution

    Run untrusted, user-submitted, and AI-generated code inside secure WebAssembly sandboxes.

    Sub-5ms cold starts

    The only platform with <5ms cold-start response times for Node.js and Python apps.

    Start Building NowView Documentation

    Try it now

    Describe an app and get a live StackMachine URL in seconds.

    Describe what you want to host, or use a quick action below. I'll deploy it on StackMachine and return a live link with a preview.

    Everything you need, all in one place

    Built-in infrastructure blocks for shipping, scaling, and operating production apps without stitching services together.

    Build

    From commit to app

    Deploy from Git

    Deployments on new commits.

    Builds

    Build any source into an Edge app.

    Domains

    Add custom domains to your apps.

    Runtime

    Stateful apps at the edge

    Serverless Compute

    Scale apps serverlessly.

    Databases

    A scalable database from day one.

    Cloud Storage

    Store and serve files.

    Operate

    Production workflows

    Analytics

    Track usage from your apps.

    Mail

    Send and track emails.

    Cron jobs

    Schedule recurring jobs.

    One deployment modelDocs for every primitiveDesigned to scale from day one

    Trusted by teams shipping production AI and hosting products

    Wasmer logo
    Simply Static logo
    Tiiny logo
    Manus logo
    ModularDS logo

    Pay only when your code runs

    Slow model calls, webhook pauses, and hibernating WebSockets should not count as compute. StackMachine meters CPU work, not wall-clock wait time.

    1ms
    Agent waits on LLM 2500ms
    0.5ms
    External API call 300ms idle
    0.75ms

    Fast global access

    Run close to users and tool calls with a network designed for short-lived, dynamic workloads.

    0req/s

    Serverless scaling

    Traffic bursts are absorbed automatically, then capacity scales down when work is complete.

    More isolated apps on the same hardware.

    Containers replicate OS and runtime state per tenant. WebAssembly shares executable binaries across isolated tenants, unlocking an order of magnitude higher compute density.

    Containers
    ~24 apps
    per server
    ~256MB per app|~500ms cold start|OS overhead per tenant
    WebAssembly
    10,000+ apps
    per server
    ~2MB per app|~5ms cold start|Shared binaries

    Each dot represents an isolated application instance. Same server. Same hardware. 400x more density.

    Source: WebAssembly Clouds: The world after Containers

    How we compare

    Run real app runtimes in secure Wasm sandboxes with sub-5ms cold starts and density built for massive scale.

    FeatureStackMachineAWS LambdaCloudflare Workers
    Cold start<5ms200ms+Varies by language (10-800ms)
    Code changes requiredNoYesYes
    Node.js supportYesYesNo
    Next.js supportYesLimitedLimited (via vinext fork)
    Python supportYesYesLimited (Pyodide)
    PHP supportYesNoNo
    WordPress supportYesNoNo
    Go supportYesYesNo
    Multithreading & multiprocessingYesYesNo
    Native realtime / WebSocketsYesNoYes
    Idle time billedNoYesNo
    Apps per serverHighest densityLower densityHigh density

    Built for teams that need reliable scale from day one.

    Whether you deploy a single agent workspace or a fleet of customer apps, StackMachine keeps the operational model simple: isolated, measurable, and fast.

    Deployments
    Requests served
    Lower compute cost
    Active regions

    Frequently Asked Questions

    Short answers to the architecture, security, and workflow questions teams ask before adopting StackMachine.

    Ship AI applications on infrastructure designed for dynamic code.

    Secure environment with real-world tools for enterprise-grade agents and hosting providers.

    Start building