generated: '2026-08-09' method: derived source: - openapi/routerplex-inference-openapi.json - well-known/routerplex-api-catalog.json - well-known/routerplex-robots.txt - mcp/routerplex-tools-list.json - https://docs.routerplex.com/errors standards: - id: openapi-3.1 conforms: true evidence: >- openapi/routerplex-inference-openapi.json declares "openapi": "3.1.0" and is served with content-type application/vnd.oai.openapi+json;version=3.1 from https://routerplex.com/openapi.json. - id: rfc9727-api-catalog conforms: true evidence: >- https://routerplex.com/.well-known/api-catalog returns a linkset with service-desc, service-doc and status link relations, content-type application/linkset+json with the RFC 9727 profile. - id: mcp-2025-06-18 conforms: true evidence: >- https://routerplex.com/mcp answers initialize with protocolVersion "2025-06-18" and tools/list with 3 tools carrying JSON Schema draft-07 inputSchemas and standard tool annotations. - id: openai-chat-completions-compatibility conforms: true evidence: >- POST /v1/chat/completions and /v1/images/generations mirror the OpenAI request shapes; docs state the official OpenAI Python and JS SDKs work unmodified with only base_url changed. - id: openai-responses-compatibility conforms: true evidence: POST /v1/responses, documented as the compatibility route for Codex custom model providers. - id: anthropic-messages-compatibility conforms: true evidence: >- POST /v1/messages accepts the Anthropic Messages shape and x-api-key header, and emits Anthropic SSE event names (message_start, content_block_delta, message_stop) for every model, not only Claude. - id: sse-streaming conforms: true evidence: >- stream:true returns server-sent events; documented at https://docs.routerplex.com/streaming. Streams stay open up to 10 minutes. - id: http-bearer-auth conforms: true evidence: >- components.securitySchemes.RouterPlexApiKey is type http / scheme bearer, applied globally via a root security requirement. - id: content-signal conforms: true evidence: >- https://routerplex.com/robots.txt publishes "Content-Signal: ai-train=no, search=yes, ai-input=yes" — an explicit machine-readable AI usage preference. - id: llmstxt conforms: true evidence: >- https://routerplex.com/llms.txt (H1 + blockquote + sectioned link lists) and https://docs.routerplex.com/llms-full.txt, plus Markdown companions at /pricing.md and /benchmarks.md. - id: rfc9457-problem-details conforms: false evidence: >- Errors use the OpenAI envelope {"error":{"message","type","code"}} with content-type application/json — not application/problem+json. - id: rfc9116-security-txt conforms: false evidence: /.well-known/security.txt returns 404 on routerplex.com, api.routerplex.com and docs.routerplex.com. - id: oauth2 conforms: false evidence: >- No OAuth. Authentication is a static API key created in the dashboard. No /.well-known/oauth-authorization-server or oauth-protected-resource. - id: oidc conforms: false evidence: /.well-known/openid-configuration returns 404 on every host. - id: rfc8594-sunset-header conforms: false evidence: No deprecation or sunset policy published; no Sunset/Deprecation headers documented. - id: a2a-agent-card conforms: false evidence: >- /.well-known/agent-card.json and /.well-known/agent.json both return 404 on routerplex.com, api.routerplex.com and docs.routerplex.com. - id: asyncapi conforms: false evidence: >- No event or webhook surface. Streaming is request-scoped SSE, not a subscribable event stream, so AsyncAPI is not applicable rather than missing. compliance_program: published: false note: >- No trust center, no SOC 2 / ISO 27001 / PCI / HIPAA / GDPR certification claims found on routerplex.com, trust.routerplex.com, or in the terms and privacy pages. The company describes itself as "young and independently built" and launched in July 2026.