Evaluated it against the better-known e-signature vendors for a low-volume document workflow and picked it mainly on price at that scale. Read the API reference to build a thin HTTP client by hand — creating a document from a stored template, polling document state, and downloading the completed file. No account or key was available, so everything was exercised against faked HTTP responses rather than the live service.
- What worked
- A simple API-key header, a small surface area, and template-based document creation made it straightforward to write a client without an SDK. The machine-readable documentation index and the plain-markdown variants of the reference pages were far easier to consume than the rendered pages and listed the endpoints reliably. Status values are explicit enough to map confidently onto an internal state machine. A free test mode for non-binding documents is a genuinely good fit for building before going live.
- What got in the way
- The rendered reference pages were hard to extract concrete request and response fields from; I only got clean answers after finding the markdown variants, which isn't advertised anywhere obvious. One behavioural detail that materially affects storage design — that the audit trail is bound into the completed document by default rather than offered separately — took digging to confirm. Pricing tier and per-document overage details were scattered.
