Skip to content
English

Switch theme

// Software stack

Voice agent stack

Low-latency speech I/O plus an actionable agent.

4 pieces4 setup stepsscenario: Creators: voice-over, podcasts and AI music

What goes in it

  1. 1

    FunASR Speech recognition & synthesis Open source

    Speech-to-text

    Industrial-grade speech recognition toolkit.

    Or: Wispr Flow · moonshine · openwhispr

    ★ 21k
  2. 2

    ElevenLabs Speech recognition & synthesis Closed / commercial

    Text-to-speech

    Lifelike speech synthesis, voice cloning and voice agent platform.

    Or: Cartesia · Fish Audio · supertonic · pocket-tts

  3. 3

    vocode-core Speech recognition & synthesis Open source

    Voice framework

    Build voice-based LLM agents — modular and open source.

    Or: dograh · speech-to-speech

    ★ 3.8k
  4. 4

    Wispr Flow Speech recognition & synthesis Closed / commercial

    Dictation

    Voice dictation in any app, cleaned up into polished text in real time.

    Or: FluidVoice · openless · Typeless

How to set it up

  1. 01Measure end-to-end latency
  2. 02Choose cloud or on-device TTS
  3. 03Add barge-in and echo cancellation
  4. 04Require confirmation for tool calls

Data updated: AI software & hot ranks · Solutions & stack guides(UTC+8)