[{"capabilities":["decision-import"],"direction":"inbound","docs_url":"./docs/agentroute-spec.md","id":"native-receipt","name":"AgentRoute receipt API","roles":["router","gateway"],"status":"available","summary":"Any application-native router can emit the versioned receipt contract directly.","transport":"JSON / JSONL"},{"capabilities":["decision-import","live-capture","policy-export"],"direction":"bidirectional","docs_url":"https://openrouter.ai/docs","id":"openrouter","name":"OpenRouter","roles":["router","gateway","policy-target"],"status":"available","summary":"Imports allowlisted routing metadata and supports explicit non-streaming capture.","transport":"response metadata / API"},{"capabilities":["decision-import","policy-export"],"direction":"bidirectional","docs_url":"https://docs.litellm.ai/","id":"litellm","name":"LiteLLM","roles":["router","gateway","policy-target"],"status":"available","summary":"Normalizes selected model and provider metadata with conservative evidence fidelity.","transport":"callback metadata"},{"capabilities":["task-pack"],"direction":"inbound","docs_url":"https://docs.exa.ai/","id":"exa","name":"Exa","roles":["task-source"],"status":"available","summary":"Builds fresh, source-linked evaluation task packs without granting routing authority.","transport":"search API"},{"capabilities":["trace-export"],"direction":"outbound","docs_url":"https://opentelemetry.io/docs/specs/semconv/","id":"opentelemetry","name":"OpenTelemetry","roles":["telemetry"],"status":"available","summary":"Exports routing-decision spans without prompts, response text, or candidate endpoints.","transport":"privacy-safe OTLP JSON"},{"capabilities":["trace-export"],"direction":"outbound","docs_url":"https://github.com/Arize-ai/openinference/tree/main/spec","id":"openinference","name":"OpenInference","roles":["telemetry"],"status":"available","summary":"Exports OpenInference LLM spans without prompts, outputs, endpoints, or arbitrary metadata.","transport":"privacy-safe OTLP JSON"},{"capabilities":["decision-import","observation-import","policy-export"],"direction":"bidirectional","docs_url":"https://portkey.ai/docs/product/ai-gateway","id":"portkey","name":"Portkey AI Gateway","roles":["router","gateway","policy-target"],"status":"available","summary":"Imports allowlisted evidence and compiles a review-only routing config artifact.","transport":"gateway logs / config export"},{"capabilities":["decision-import","observation-import","policy-export"],"direction":"bidirectional","docs_url":"https://vercel.com/docs/ai-gateway","id":"vercel-ai-gateway","name":"Vercel AI Gateway","roles":["router","gateway","policy-target"],"status":"available","summary":"Imports routing evidence and compiles review-only AI SDK gateway provider options.","transport":"provider metadata / provider options"},{"capabilities":["decision-import","observation-import"],"direction":"inbound","docs_url":"https://developers.cloudflare.com/ai-gateway/","id":"cloudflare-ai-gateway","name":"Cloudflare AI Gateway","roles":["gateway","telemetry"],"status":"available","summary":"Imports documented log fields as routing decisions and measured operational observations.","transport":"logs / analytics / OpenTelemetry"},{"capabilities":["evaluation-import"],"direction":"inbound","docs_url":"https://www.braintrust.dev/docs/evaluate","id":"braintrust","name":"Braintrust","roles":["evaluator","telemetry"],"status":"available","summary":"Imports numeric experiment and online scores without retaining prompts, outputs, or reasoning.","transport":"experiment / trace results"}]