NudgeHost developer resources
Everything developers and agents need to work with NudgeHost, linked from one page. The REST API, the MCP server, and the discovery documents that let software find all of it without being told where to look.
https://www.nudgehost.com/api/mcp
Stateless streamable HTTP with four tools, share_text, create_upload_url, complete_upload and list_links. It accepts POST and takes a bearer token, either an API key created from the dashboard or an OAuth access token. An unauthenticated call answers 401 with a WWW-Authenticate challenge naming the metadata document below. The API documentation covers the request shape for each tool.
- Agent skillJSONDiscovery index naming the nudgehost skill, where to fetch it, and a SHA-256 digest of its bytes./.well-known/agent-skills/index.json
- Agent identityJSONAID v2 record naming the MCP endpoint, its auth scheme, and the documentation. The TXT record at _agent.nudgehost.com carries the same thing./.well-known/agent
- API catalogueLinksetThe RFC 9727 catalogue, a linkset pointing at the OpenAPI specification and the documentation page./.well-known/api-catalog
- OAuth metadataJSONRFC 9728 protected-resource metadata naming the authorization server that issues access tokens for the API./.well-known/oauth-protected-resource
- Plain-text summaryTextWhat the service does, how agents authenticate, and where every document on this page lives./llms.txt
- Markdown for agentsMarkdownEvery page that is also served as markdown, grouped by section, with the markdown URL for each./sitemap.md
Use the www hostname
The apex domain nudgehost.com answers with a 308 to www.nudgehost.com. That is a change of origin, so a client following it drops the Authorization header on the way and the request arrives with no credentials. Send API calls to www.nudgehost.com.
Marketing pages are also served as markdown
Every page in the host, viewers, dev-tools, use-cases, features, compare, glossary and blog sections returns markdown to a request whose Accept header asks for text/markdown, at the same URL a browser gets HTML from. The same documents are reachable without content negotiation by adding /index.md to the path, and the markdown sitemap lists every one of them.