S3-Compatible Gateway for Arweave Settlement
Use SigV4 S3 APIs against Evercore: logical buckets per API key map to encrypted file operations and permanent Arweave artifacts — reuse AWS CLI and SDKs without a second storage silo.
Use PutObject, GetObject, ListObjects, and SigV4 the way your toolchain already does. Buckets are logical namespaces tied to your API key; every object still flows through encryption, versioning, and permanent settlement.
The storage API the world already integrated — with permanence underneath.
Buckets are logical namespaces in Evercore’s control plane, not third-party cloud buckets. SigV4-authenticated S3 calls map into the same encryption, versioning, and settlement pipeline you already use over REST — so lift-and-shift tools, CLIs, and ETL jobs can land without a rewrite.
Value
Faster procurement and engineering sign-off
Teams that already approved “S3-shaped” storage can evaluate Evercore as a drop-in interface change instead of a new storage paradigm.
One identity model, two surfaces
Your Evercore API key owns buckets and scopes; optional S3 access keys exist only to satisfy SigV4 — they do not create a parallel security universe.
Permanent artifacts, familiar operations
PutObject and friends still settle to encrypted Arweave transactions. You keep chain-of-custody, receipts, and governance hooks from the rest of the platform.
How it works
Evercore API key
Your primary identity: scopes, billing, and ownership of every bucket row. Created in Console or admin APIs.
S3 access key + secret (optional)
Generated for SigV4 only. The secret never replaces your API key — it signs HTTP requests the AWS SDK expects.
Bucket record
A logical namespace in Evercore’s database (name + owner_key_id), not an AWS account. Required before multipart and presigned flows.
S3 object key
Path-style keys become Evercore file names. Metadata headers map to Arweave tags through the gateway mapper.
For operators
- Create buckets through the gateway or the console admin API so multipart, presigned flows, and listings have a consistent namespace anchor.
- Issue S3 credentials per API key when tooling requires HMAC-SHA256 signing; rotate or revoke them without changing your core key’s scopes.
- Point aws-cli or SDKs at the Evercore S3 endpoint and keep object keys path-shaped — they map directly to Evercore file names for retrieval and tagging.
For buyers
- Closes the “we only support S3” objection common in regulated data platforms and legacy ingest.
- Reduces migration cost compared with bespoke Arweave adapters for every language and runtime.
- Pairs with Professional+ plans where gateway features are enabled — align sales narrative with plan gates in docs and console.
Is this Amazon S3?
No. It is an S3-shaped HTTP API. Buckets live in Evercore’s control plane; objects settle to Arweave through Evercore’s processing pipeline. You get compatibility, not AWS infrastructure.
Why create a bucket if Arweave has no buckets?
Buckets isolate namespaces per API key, drive S3-compatible listings, and anchor multipart and presigned upload state. They are metadata — the bytes still land as tagged transactions.
Can I skip S3 and use REST only?
Yes. S3 is optional. Teams that already run AWS SDKs or Spark connectors can adopt faster; greenfield apps can stay on the JSON APIs.