Roadmap

0.1.0 is published. This page lists known gaps; the changelog lists implemented changes.

Next

• Run ShipStation v2 purchase and void against an account approved for safe testing. Rating is covered live; those two paths are the largest untested money paths in the library.
• Add a live ShipStation v1 testLabel purchase test. The retained no-charge evidence covers one Stamps.com/USPS label and is not generalised to other carriers.
• Populate per-package label output. Label.parcel_labels was removed rather than shipped permanently empty, and implementing it needs a captured v2 purchase response.
• Use Shippo's native multi-piece rating where the carrier and account support it, instead of always summing per-parcel quotes.

Adapter work

Sendcloud

Sendcloud is the next candidate because it adds another aggregator shape. Before an adapter is added, verify:

• service and carrier identifiers;
• multi-parcel request and purchase behavior;
• customs value basis and duty fields;
• test credentials and purchase safety;
• label cancellation/refund semantics;
• manifest requirements.

Direct carriers

Direct USPS, UPS or FedEx adapters would test whether the adapter contract assumes an aggregator account model. They require separate work on authentication, account configuration and manifests.

Later

• tracking normalization across providers;
• manifest and close-out operations;
• address validation;
• connection-pooled optional transports;
• recorded HTTP replay tooling for adapter development;

Outside this project

shipzil does not provide provider health scoring, automatic service equivalence, cost optimization or automatic purchase failover. Those decisions need separate policy and evidence. The Gateway keeps provider-specific code out of an application; it does not remove the operational need for multiple funded provider accounts.