FAQ for engineers

Engineer-focused answers. For plan pricing questions see Plans & billing.

​

Can I self-host BlobRouter?

Not as a supported product today. The control plane targets Cloudflare Workers, Queues, and KV. The monorepo is readable for audit and contribution; a packaged self-host story is not v1.

​

Why not just use S3 lifecycle policies?

Lifecycle rules are useful and we are not anti-lifecycle. BlobRouter adds: cross-provider economics (R2/B2), a read-only waste report and decision report before you change anything, explicit migration approval, and ongoing routing for new uploads — not only aging objects inside one AWS account.

​

Why not S3 Intelligent-Tiering?

Intelligent-Tiering optimizes within S3. BlobRouter compares across providers (including zero-egress R2 and cheap B2 storage) and shows estimated savings vs staying on S3. Many teams use both: Intelligent-Tiering inside AWS plus BlobRouter for audit and multi-provider routing.

​

Can I bring my own bucket?

Yes. That is the default. You own the buckets; BlobRouter stores encrypted credentials (router) or assumes your role (scanner) and never hosts objects.

​

Can I use only the scanner?

Yes. Connect AWS, run the free audit, ignore the SDK. Migration and routing are optional next steps. Continuous analysis on paid plans still applies to the estate under audit.

​

What is the decision report vs recommendations?

Recommendations are the legacy per-object / per-bucket migrate suggestions you can approve. The decision report is the Storage Decision Engine view (callouts, segment economics, scenarios, DO NOT MOVE / monitor). It is additive on scan results when present — older scans may omit it. See Storage Decision Engine.

​

Is the marketing demo the same as my live audit?

No. The site demo uses a representative fixture workload so the full report UI is populated. Live audits start from your observed S3 data and label modeled assumptions. Do not treat demo dollars as your production savings.

​

Can I disable routing?

Do not call the upload APIs / SDK. Scanner and dashboard audit continue to work. There is no separate “routing off” flag required.

​

How is billing calculated?

You pay for continuous analysis of the estate on paid plans — not for migrations. Separately, GB routed = bytes that complete a BlobRouter routing decision (usage_events for uploads); audits and analysis do not count toward that quota. See Plans.

​

Can I use private networking?

Uploads go from your environment to the provider endpoint (public S3-compatible APIs today). PrivateLink-style paths are not a v1 feature. Your apps can still run in private networks as long as they can reach the provider and api.blobrouter.com.

​

Do uploads pass through BlobRouter?

No. Only control-plane JSON (init/complete) hits BlobRouter. Bytes go client → provider via presigned URL. See Request lifecycle.

​

Can I integrate with an existing S3 SDK?

Yes for direct provider access you already have. BlobRouter’s SDK talks to the control plane, then uses HTTPS PUT to the presigned URL. You can also call the REST API from any language and PUT yourself.

​

Is the SDK safe in the browser?

No for v1. Keep API keys on the server (Node, Python workers, backends). Browser upload would require a different key model.

​

Where do I find why you built it this way?

Architecture decisions.