anvil-cloud

Cloud infrastructure that's secure by default — not by accident.

Anvil wraps raw cloud resources into opinionated, production-ready components. No boilerplate. No copy-pasting security configs. Just declare what you need.

Built on Pulumi.

Install

pip install anvil-cloud

Secure by default

Every Anvil component ships with defaults aligned to production from day one — public access blocked, encryption enforced, cost tags applied. The goal isn't to make compliance automatic, but to make it a platform you can actually build on.

The App class

Every Anvil program starts with anvil.App(). The run callback receives a Context with:

• ctx.stage — current deployment stage (defaults to your OS username for dev isolation)
• ctx.project — project name from anvil.yaml
• ctx.export(name, value) — export stack outputs
• ctx.providers — named cloud providers for multi-region / multi-account
• ctx.ref(id) — reference a resource declared elsewhere (see below)

Forward references (ctx.ref)

Normally you reference a resource through the variable it's assigned to, which means it must be declared first. ctx.ref("id") lets you reference a resource by its logical name — even one declared later in the program:

def infra(ctx: anvil.Context):
    fn = anvil.aws.Lambda("processor",
        runtime=LambdaRuntime.NODEJS20_X,
        handler="index.handler",
        environment={
            # forward reference — `db` is declared below
            "DSQL_ENDPOINT": ctx.ref("db").endpoint,
        },
    )

    db = anvil.aws.DSQL("db", schemas=[...])

ctx.ref("db").endpoint resolves to the same output as db.endpoint — they're interchangeable.

This is an escape hatch for cases where declaration order is awkward; most code references resources directly. Two things to know:

• Keep references flowing one direction. Because you can point either way, you can accidentally create a cycle (A references B's output and B references A's). That deadlocks the deploy rather than erroring cleanly — so don't do it.
• A typo or never-declared id fails with a clear error before anything deploys.

Running locally: anvil deploy uses whatever python3 is on your PATH, so activate your project venv first (source .venv/bin/activate) or the program can't import pulumi.

Grants

Grants are how Anvil wires permissions between resources. Instead of writing IAM policies by hand, you call .grant() on a resource and Anvil handles both the IAM role policy and the environment variable injection automatically.

A Lambda reading from a Bucket:

import anvil_cloud as anvil

def infra(ctx: anvil.Context):
    bucket = anvil.aws.Bucket("uploads",
        data_classification="sensitive",
    )

    fn = anvil.aws.Lambda("processor",
        runtime="nodejs20.x",
        handler="index.handler",
        code="./src",
    )

    # Grants the Lambda read access to the bucket and scopes down to specific bucket paths.
    # Anvil creates the IAM policy and injects UPLOADS_BUCKET_NAME
    # into the Lambda's environment automatically.
    bucket.grant(fn, actions=["read"], path=["user/*"])

anvil.App(run=infra)

What Anvil does under the hood:

• Creates an IAM RolePolicy scoped to the exact actions requested
• Injects the resource identifier as an environment variable on the target (e.g. UPLOADS_BUCKET_NAME)
• No manual ARN wiring, no forgotten permissions

SvelteKit deployment

Deploy a SvelteKit app to AWS with a single component. Anvil provisions S3, CloudFront, ACM, Lambda (via Lambda Web Adapter), and Route53 — with HTTPS and a custom domain out of the box:

import anvil_cloud as anvil

def infra(ctx: anvil.Context):
    site = anvil.aws.SvelteKitSite("web",
        domain="myapp.com",
    )
    ctx.export("url", site.url)

anvil.App(run=infra)

Overrides

Every component accepts a transform argument to override the underlying resource config when you need to break from the defaults:

bucket = anvil.aws.Bucket("custom",
    data_classification="non-sensitive",
    transform=anvil.aws.BucketTransformArgsArgs(
        overrides=anvil.aws.BucketOverridesArgs(
            force_destroy=True,
            tags={"env": "dev"},
        ),
    ),
)

Requirements

• Python >= 3.8
• Pulumi >= 3.0.0
• Anvil CLI: curl -fsSL https://raw.githubusercontent.com/DamienPace15/anvil/master/install.sh | sh

Links

• Docs
• GitHub
• npm SDK
• Go SDK

License

Apache-2.0