generated: '2026-07-21' method: derived source: openapi/, asyncapi/, https://docs.slng.ai/protocols.md, https://docs.slng.ai/region-override.md authentication: style: bearer-token header: 'Authorization: Bearer ' note: >- Opaque API key issued from the SLNG dashboard. Some catalog (read) endpoints accept anonymous requests; an API key personalizes pricing and region availability. cross_ref: authentication/slng-authentication.yml protocols: transports: [https, wss] note: >- Every model can be invoked over HTTP (request/response) or WebSocket (wss) for streaming STT/TTS. See asyncapi/ for the streaming surface. pagination: style: page-number params: [page, page_size] response_fields: [meta.page, meta.page_size, meta.total, meta.pages] defaults: {page: 1, page_size: 20, max_page_size: 100} regional_routing: headers: - name: X-Region-Override description: Force a specific region (e.g. eu-north-1, us-east-1, asia-south1). Auto-selected if omitted. - name: X-World-Part-Override description: Force a world part (e.g. eu, ap, au). Auto-selected if omitted. note: SLNG orchestrates compute across 60+ regions / 11 sovereign hubs for latency and data residency. request_tracking: fields: - tag (STT requests) — free-form request-tracking tag - metadata (batch jobs) — arbitrary key-value metadata attached to a job response_ids: metadata.request_id is returned on transcription responses idempotency: supported: partial note: >- Batch jobs accept a client-supplied alphanumeric job_id (auto-generated if omitted). No general Idempotency-Key header is documented for the synchronous STT/TTS surface. versioning: scheme: uri-path current: v1 error_envelope: shape: '{"error": string}' provider_shape: '{"error": string, "status": int, "details": string}' cross_ref: errors/slng-problem-types.yml