Every site published with RELAYTO, on a *.relayto.com address or on the publisher's own domain, has a public, read-only REST API and MCP server. No API key and no sign-in: they return exactly what any visitor can open without signing in. Replace <site> below with the site's host name, for example benefits.relayto.com.
Endpoints at a glance
| What | Where |
|---|---|
| REST API | https://<site>/api/site/v1 |
| OpenAPI spec | https://<site>/openapi.json |
| API reference | https://<site>/developers |
| MCP server | https://<site>/mcp (tools: search_content, list_hubs, get_content) |
| For LLMs | https://<site>/llms.txt, and any page as Markdown at <page URL>.md |
REST API endpoints
GET /api/site/v1/: the site and its API linksGET /api/site/v1/search?q=…: search pages, hubs and documentsGET /api/site/v1/pages: the site's own pagesGET /api/site/v1/hubs,/hubs/{hubSlug},/hubs/{hubSlug}/items: hubs and what they containGET /api/site/v1/contents,/contents/{contentSlug}: documents, with the full text of every page as MarkdownPOST /api/site/v1/exports,GET /api/site/v1/exports/{exportId}: Markdown of up to 10 documents at once
Lists are paginated with cursors; errors use application/problem+json. The full reference, with rate limits and error codes, is on each site's /developers page.
Try it
curl "https://benefits.relayto.com/api/site/v1/search?q=open%20enrollment"
npx relayto search "open enrollment" --site benefits.relayto.com
Use a site from an AI agent
claude mcp add --transport http benefits https://benefits.relayto.com/mcp
For clients that only start local servers:
npx -y relayto-mcp --site benefits.relayto.com
All sites are listed in the official MCP Registry as com.relayto/site-content.
What stays private
Content behind a password, a sign-in or an email form is never returned in full. Private content answers 403 with code not_public; email-gated content returns its description only.