[[WRITE: voice pipeline headline]]

[[WRITE: intro — generic real-time voice pipeline, anchored on RegalaUnaLlamada]]

  1. Audio capture

    [[WRITE: how audio is captured / streamed]]

  2. Speech-to-text (STT)

    [[WRITE: streaming STT approach, latency budget]]

  3. LLM reasoning

    [[WRITE: prompt/agent orchestration, tool use]]

  4. Text-to-speech (TTS)

    [[WRITE: streaming TTS, barge-in handling]]

  5. Realtime transport

    [[WRITE: transport, turn-taking, latency]]