generated: '2026-09-14' method: searched source: https://docs.aiola.ai/developer-guides/text-to-speech/quickstart, https://docs.aiola.ai/developer-guides/speech-to-text/quickstart limit_count: 0 rate_limits: [] headers: [] status_on_exhaustion: undocumented note: >- aiOla publishes no numeric rate limit. The full extent of the documented guidance is one bullet on the text-to-speech quickstart — "Rate Limiting: Be mindful of API rate limits for production applications" — with no scope, window, ceiling, burst allowance, response header, or exhaustion status code. The limits clearly exist: aiOla's own SDK ships RATE_LIMIT_EXCEEDED and MAX_CONCURRENCY_REACHED as error codes it expects the service to return. An honest zero is recorded rather than a guess. observed_constraints: - constraint: maximum audio file size value: 50 MB scope: per request, POST /api/speech-to-text/file source: https://docs.aiola.ai/developer-guides/speech-to-text/quickstart note: >- A payload cap, not a rate limit — recorded here because it is the only published numeric ceiling on the API and it is what a batch client will hit first. Larger audio must be streamed or split. - constraint: concurrent sessions per account value: undocumented scope: per account signal: error code MAX_CONCURRENCY_REACHED release: DELETE https://auth.aiola.ai/voip-auth/session note: >- The binding limit on this API is almost certainly concurrency rather than requests-per-second, since a streaming session holds a slot for its whole lifetime. The ceiling is not published, so a caller cannot size a workload before running it. - constraint: access token lifetime value: ~30 minutes scope: per session source: https://docs.aiola.ai/developer-guides/authentication gaps: - No RateLimit-* / X-RateLimit-* response headers are documented. - No Retry-After behaviour is documented. - No status code is documented for an exhausted limit. - The concurrency ceiling — the limit the error taxonomy implies matters most — has no published value.