Realtime

Realtime WebSocket

OpenAI-compatible Realtime WebSocket for voice, multimodal, and transcription sessions.

Connect with wss://gen.pollinations.ai/realtime?model=openai/gpt-realtime-2.1 and send/receive OpenAI Realtime JSON events over the socket. Selecting elevenlabs/scribe-v2-realtime creates a transcription session automatically. Server clients can authenticate with Authorization: Bearer <key>. Browser WebSocket clients can use ?key=pk_... because they cannot set custom authorization headers.

Models: microsoft/mai-transcribe-2-streaming, openai/gpt-realtime-2.1, openai/gpt-realtime-2.1-mini, elevenlabs/scribe-v2-realtime, openai/gpt-live-transcribe.

Billing: requires a positive balance and settles one session total when the socket closes.

get/realtime

Request

  • Base URL: https://gen.pollinations.ai
  • URL: https://gen.pollinations.ai/realtime
  • Auth: HTTP bearer

Query parameters

model'microsoft/mai-transcribe-2-streaming' | 'openai/gpt-realtime-2.1' | 'gpt-realtime-2.1' | 'gpt-realtime-2' | 'openai/gpt-realtime-2.1-mini' | 'gpt-realtime-2.1-mini' | 'elevenlabs/scribe-v2-realtime' | 'scribe-realtime' | 'openai/gpt-live-transcribe' | 'gpt-live-transcribe'

Realtime model to use. Supported models: microsoft/mai-transcribe-2-streaming, openai/gpt-realtime-2.1, openai/gpt-realtime-2.1-mini, elevenlabs/scribe-v2-realtime, openai/gpt-live-transcribe.

keystring

Pollinations API key. Useful for browser WebSocket clients that cannot set custom Authorization headers.

Changes

No recorded changes to this operation across all 1 revision of this API.