railway.com publishes an Agentic Resource Discovery manifest at
/.well-known/ard.json, declaring what an AI agent can call on this domain.
Everything below is read from that file. First seen by our crawler on 31 August 2026.
| Declared resource | Type |
|---|---|
| Authenticating to Railwayanswers in 166ms The four ways an agent or automated client holds a Railway credential — CLI login, remote MCP over OAuth, API tokens, and Login with Railway — with discovery, claim, error, and revocation walkthroughs. endpoint: https://railway.com/auth.mdidentifier: urn:air:railway.com:doc:authpublished to be found for: how does an agent authenticate to Railway · how do I get a Railway API token docs auth trust identity: {"principalName": "railway.com"} | text/markdown |
| Railway API reliabilitynot yet checked How the Railway API behaves under failure: the error convention and code catalog, rate limits and headers, retry/idempotency guidance, and the versioning/deprecation policy. endpoint: https://railway.com/api-reliability.mdidentifier: urn:air:railway.com:doc:api-reliabilitypublished to be found for: what are the Railway API rate limits · how does the Railway API report errors · is the Railway API safe to retry docs api reliability trust identity: {"principalName": "railway.com"} | text/markdown |
| Railway for AI Agentsnot yet checked Setup briefing for coding agents: CLI install, local MCP and agent skills, remote MCP, and the core deploy/database/volume workflows. endpoint: https://railway.com/agents.mdidentifier: urn:air:railway.com:doc:agents-briefingpublished to be found for: how do I set up Railway as a coding agent · install the Railway CLI for an AI agent docs agents trust identity: {"principalName": "railway.com"} | text/markdown |
| Railway llms.txtnot yet checked Navigation index of Railway's agent-readable pages: product, pricing, docs, templates, changelog, and machine-readable endpoints. endpoint: https://railway.com/llms.txtidentifier: urn:air:railway.com:doc:llms-txtpublished to be found for: what is Railway · where are Railway's agent-readable docs docs index trust identity: {"principalName": "railway.com"} | text/plain |
| Railway Remote MCP Serveranswers in 152ms Deploy and operate apps, databases, volumes, and cron jobs on Railway from any MCP client. endpoint: https://railway.com/.well-known/mcp/server-card.jsonidentifier: urn:air:railway.com:server:railway-mcppublished to be found for: deploy my app to Railway · add a Postgres database to my Railway project · show me the logs for my failing Railway deployment · set an environment variable on my Railway service mcp deploy infrastructure trust identity: {"principalName": "railway.com"} | application/mcp-server-card+json |
| Railway OpenAPI descriptionnot yet checked OpenAPI 3.1 description of Railway's HTTP surface: the GraphQL entrypoint with its bearer/OAuth2 security and named scopes, plus the OAuth 2.0 / OIDC endpoints. endpoint: https://railway.com/openapi.jsonidentifier: urn:air:railway.com:api:openapipublished to be found for: what are Railway's API endpoints · how do I authenticate to the Railway API api openapi trust identity: {"principalName": "railway.com"} | application/openapi+json |
| Railway API catalog (RFC 9727)not yet checked Linkset naming every Railway API, its documentation, its auth metadata, and its status page. endpoint: https://railway.com/.well-known/api-catalogidentifier: urn:air:railway.com:catalog:api-catalogcatalog api trust identity: {"principalName": "railway.com"} | application/linkset+json |
| Railway Public API (GraphQL)answers in 229ms Railway's public API — a single GraphQL endpoint with the schema available by introspection. Authentication and token types are described at https://railway.com/auth.md. endpoint: https://backboard.railway.com/graphql/v2identifier: urn:air:railway.com:api:public-graphqlpublished to be found for: list my Railway projects via the API · trigger a Railway deployment from CI api graphql trust identity: {"principalName": "railway.com"} | application/graphql-response+json |
https://railway.com/.well-known/ard.json, this publisher's own file.
Identifiers are reproduced exactly as published.
Of the resources listed, 3 have been probed for reachability.
"Answers" and "no response" describe only whether an endpoint replied when last probed:
services go down and come back, and an entry is demoted rather than removed. None of this
is a trust, quality or safety rating.
https://neuronto.com/search, or connect an agent to https://neuronto.com/mcp.
If this is your domain, the free console shows what a registry
sees when it reads your manifest.