Skip to main content
Kokoro is a local Text-to-Speech (TTS) engine. It generates lifelike voice audio directly on your machine, without requiring an internet connection or external API. Designed for low-latency and real-time use cases, Kokoro is ideal for projects that need fast, offline voice synthesis—such as on-device agents, prototypes, or privacy-sensitive applications. The KokoroTTS plugin in the Vision Agents SDK allows you to use the Kokoro model with configurable voices, playback speed, and device support (CPU/GPU).

​
Installation

Install the Stream Kokoro plugin with 
uv add tream-agents[kokoro]

​
Example

Check out our Kokoro example to see a practical implementation of the plugin and get inspiration for your own projects, or read on for some key details.

​
Initialisation

The Kokoro plugin for Stream exists in the form of the TTS class: 
from vision_agents.plugins import kokoro

tts = kokoro.TTS()
To initialise without passing in the API key, make sure the KOKORO_API_KEY is available as an environment variable. You can do this either by defining it in a .env file or exporting it directly in your terminal.

​
Parameters

These are the parameters available in the KokoroTTS plugin for you to customise:
NameTypeDefaultDescription
lang_codestr"a"Language code for the TTS model. "a" refers to American English.
voicestr"af_heart"The voice style or speaker preset to use.
speedfloat1.0Playback speed multiplier. Use values like 0.9 (slower) or 1.2 (faster).
sample_rateint24000Audio sample rate in Hz.
devicestr or NoneNoneThe device to run synthesis on (e.g., "cuda" or "cpu"). Auto-detected if not specified.

​
Functionality

​
Send text to convert to speech

The send() method sends the text passed in for the service to synthesize. The resulting audio is then played through the configured output track. 
tts.send("Demo text you want AI voice to say")
⌘I