What causes "no aws provider found" in anomalyco/sst?

Ensure sst.config.ts defines an AWS provider (explicit sst provider or implicit via AWS components) and redeploy

no aws provider found

Error message

no aws provider found

What it means

AwsResource.config() requires an 'aws' provider to be present in the SST project; this error is thrown when project.Provider("aws") finds none. Any AWS-backed resource (bucket files, distributions, function code updates, KV, etc.) calling config() during Create/Update/Delete fails. It means the app's configuration does not define an AWS provider, so no AWS credentials/region can be resolved.

Source

Thrown at pkg/server/resource/resource.go:60

type UpdateResult[T any] struct {
	Outs T `json:"outs"`
}

type DeleteInput[T any] struct {
	ID   string `json:"id"`
	Outs T      `json:"outs"`
}

type AwsResource struct {
	context context.Context
	project *project.Project
}

func (a *AwsResource) config() (aws.Config, error) {
	result, ok := a.project.Provider("aws")
	if !ok {
		return aws.Config{}, fmt.Errorf("no aws provider found")
	}
	casted := result.(*provider.AwsProvider)
	return casted.Config(), nil
}

func Register(ctx context.Context, p *project.Project, r *rpc.Server) error {
	awsResource := &AwsResource{ctx, p}
	cloudflareResource := &CloudflareResource{ctx, p}
	vercelResource := &VercelResource{ctx, p}
	r.RegisterName("Resource.Run", NewRun())
	
	// AWS Resources
	r.RegisterName("Resource.Aws.BucketFiles", &BucketFiles{awsResource})
	r.RegisterName("Resource.Aws.DistributionDeploymentWaiter", &DistributionDeploymentWaiter{awsResource})
	r.RegisterName("Resource.Aws.DistributionInvalidation", &DistributionInvalidation{awsResource})
	r.RegisterName("Resource.Aws.FunctionCodeUpdater", &FunctionCodeUpdater{awsResource})
	r.RegisterName("Resource.Aws.FunctionEnvironmentUpdate", &FunctionEnvironmentUpdate{awsResource})
	r.RegisterName("Resource.Aws.HostedZoneLookup", &HostedZoneLookup{awsResource})

View on GitHub (pinned to a0bd20f762)

Solutions

Ensure sst.config.ts defines an AWS provider (explicit sst provider or implicit via AWS components) and redeploy
Set AWS credentials (AWS_PROFILE, AWS_ACCESS_KEY_ID/SECRET, or SSO login) so the provider initializes
If the resource is no longer wanted, remove it from the app/state instead of updating it
Check that the provider is not renamed away from the "aws" key

Example fix

// before (no aws provider)
// sst.config.ts only has cloudflare provider
// after
const awsProvider = new providerAws("aws", { region: "us-east-1" });

Defensive patterns

Strategy: validation

Validate before calling

// in sst.config.ts, before using AWS components
if (!app.providers?.aws && !process.env.AWS_PROFILE && !process.env.AWS_ACCESS_KEY_ID) {
  throw new Error("Define an aws provider in sst.config.ts or set AWS credentials");
}

Try / catch

if err := resource.Create(ctx, input); err != nil {
    if strings.Contains(err.Error(), "no aws provider found") {
        return fmt.Errorf("app config has no aws provider; add one to sst.config.ts: %w", err)
    }
    return err
}

Prevention

Always declare an aws provider (explicit or via an AWS component) before using Resource.Aws.* resources
Keep the provider name literally "aws"
Run `sst deploy` after config changes so providers are re-initialized
Log in with AWS SSO / set credentials before deploying

When it happens

Trigger: A resource registered under Resource.Aws.* is created/updated/deleted while the project has no provider named "aws" — e.g. sst.config.ts never creates provider.aws or the app targets another cloud only.

Common situations: Using an AWS component without an implicit/default AWS provider in a config that removed it; renaming the provider so it's no longer keyed "aws"; running against a Cloudflare/Vercel-only app; stale state referencing AWS resources after switching clouds.

Related errors

AI-assisted analysis of anomalyco/sst@a0bd20f762 (2026-08-30). Data as JSON: /api/errors/3d8e766cc3880c63. Report an issue: GitHub.