Machine-readable surfaces
DIG is built to be driven by agents and tools, not just read by humans. Every contract has a stable, documented, machine-readable form so you can integrate without scraping prose.
At the docs site root
| Surface | What it is |
|---|---|
/openrpc.json | OpenRPC 1.2.6 document for the dig JSON-RPC read interface — methods, request/response JSON Schemas, and catalogued error responses. Generated from the same source as the Methods page. |
/error-codes.json | The cross-surface error catalog: dig RPC -32xxx codes, digstore CLI exit codes, DIGHUb codes, and dig-loader codes, as [{surface, code, http_or_exit, description}]. |
/llms.txt | A link-rich markdown map of the docs (llms.txt convention). |
/knowledge-graph.json | Entities (concepts + docs) and typed edges (defines, part-of, requires, see-also). |
/sitemap.xml | Every public route. |
/robots.txt | Indexing policy (public docs — indexing invited). |
Every doc page also carries schema.org JSON-LD (TechArticle / DefinedTerm), and the site emits an Organization + WebSite @graph with a SearchAction and Dataset pointers to /openrpc.json and /error-codes.json.
The hub /v1 control plane
The REST control plane (hub.dig.net/v1, bearer-JWT) — accounts, stores, domains, teams, analytics, webhooks — is described by OpenAPI served by the hub itself, with a JWKS endpoint for verifying tokens and a live pricing endpoint for the current uniform capsule price (USD-pegged $DIG). Use those endpoints to manage stores programmatically; the read path is the separate dig RPC above.
Discovery affordances
- The
digstoreCLI ships--help-json,completion <shell>, and catalogued exit codes — introspect it without out-of-band knowledge. - The dig RPC is self-describing via the OpenRPC document (method listing + schemas).
@dignetwork/dig-sdkexposescapabilities()so you can branch on what's available. → The DIG SDK
Related
- Error codes — branch on the code, never the prose
- Methods — the dig RPC method set
- The DIG SDK — the typed integration surface