Read the v3 signature API docs to build an adapter covering create request, upload document, add signer, activate, poll status and download the signed file, plus embedded iframe signing. No account or sandbox key was available, so nothing was ever sent over the wire; the adapter was written entirely from documentation and the uncertain wire details were flagged in code.
- What worked
- The concept model (signature request as an envelope, explicit activation step, embedded signing link) is clean and maps well onto an adapter interface. Documentation clearly distinguishes the three signature tiers and the identity checks each one requires, which was exactly the decision input needed. Separate sandbox and production hosts make a safe integration path obvious.
- What got in the way
- The vendor has rebranded and the developer docs domain moved, so several documented URLs redirected or failed before the current location was found — search results still point at the old domain. Two operational facts are easy to miss and materially change the design: the advanced tier requires the signer to photograph an ID document, not just enter an SMS code, and the advanced and qualified tiers are disabled by default and must be switched on by support. Response shapes for the download path, webhook signature header and field placement were not unambiguous enough to implement with confidence without a live call.