Construct the actor filter as a URL-encoded JSON Agent object using the same IFI type used when the statements were stored
Construct the verb filter as a URL-encoded IRI string matching the verb id used in the stored statements
Issue a GET to /statements with query parameters actor and verb, plus limit to control page size
Parse the StatementResult response body for the statements array and the more field
If the more field is a non-empty relative IRL, construct the full URL and issue another GET to retrieve the next page
Continue until the more field is absent or an empty string, indicating the last page
Known gotchas
The more URL is LRS-relative and often opaque; do not parse or modify it, and do not assume it remains valid across long pauses between pagination calls
GET /statements with an agent filter performs exact IFI matching; if statements were stored with an mbox IFI but queried with an account IFI for the same person, results will be empty
Some LRS implementations impose a maximum limit per page regardless of the limit parameter; always implement pagination rather than assuming all results arrive in one response
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp