API protocol

Audio, transcription, and Realtime API

Direct answerSpeech synthesis, transcription and realtime sessions differ in inputs, response formats and connection lifecycle. Audio streams and WebSockets cannot use ordinary chat JSON success checks. The current catalog provides route references, not complete audio or realtime examples.

Updated · Reviewed

Scope

Speech synthesis, transcription and realtime sessions differ in inputs, response formats and connection lifecycle. Audio streams and WebSockets cannot use ordinary chat JSON success checks. The current catalog provides route references, not complete audio or realtime examples.

Select an endpoint and model

The marketplace owns exact model IDs, endpoint support, group access and prices. Similar model names do not establish identical parameters or permissions. Provider documentation explains capabilities, not whether a model is enabled or priced on this site. Complete examples and route-only references are labeled separately.

Prepare credentials and requests

Manage model keys on the Keys page and account credentials in Profile. Browser sessions, model keys and personal access tokens serve different purposes. Examples read placeholder credentials from server environment variables and contain no real account data. Confirm parameters for the chosen endpoint and channel; example values are not defaults. The API origin is https://api.openfox.vip.

Validate results and handle failures

Check HTTP status and business results together. Text requires complete output, images require actual decoding, and tasks require a successful terminal state and valid artifacts. A task identifier alone is not completion. Bound read-only retries and never automatically repeat creation. After a timeout preserve the request ID and inspect usage logs to avoid duplicate work or charges.

Pricing and maintenance

This guide does not promise fixed prices, quotas, retention periods or provider discounts. Use current marketplace prices and account permissions. An administrator must confirm new prices or unlisted models. Examples are generated from the OpenFox OpenAPI definitions; contract changes must update fields, result checks and failure handling together. Source review is not upstream business verification; this migration makes no paid model calls.

Use cases

  • Synthesize and stream speech
  • Transcribe or translate uploaded audio
  • Build a low-latency bidirectional voice session

API protocols

  • /v1/audio/transcriptions
  • /v1/audio/translations
  • /v1/audio/speech
  • /v1/realtime

FAQ

Where can I confirm models and prices?

Use the current marketplace model IDs, endpoints and group prices.

Can I reuse another gateway endpoint directly?

No. Verify OpenFox routes, plugin contracts and authentication requirements.

Is source review a successful business test?

No. No paid calls were made; upstream availability depends on channels and permissions.

Official sources

  1. OpenAI Audio and Speech Guide Official
  2. OpenAI Realtime Guide Official
  3. OpenAI Audio API Reference Official
  4. OpenFox API Official