Speechify

Speechify provides high-quality, low-latency text-to-speech. It’s a great fit for real-time voice applications such as voice AI agents and conversational interfaces.

Vision Agents uses Stream Video for real-time WebRTC transport by default. External WebRTC transports are supported as well. Most AI providers offer free tiers to get started.

Get your Speechify API key from the Speechify console.

Installation

uv add "vision-agents[speechify]"

Quick Start

from vision_agents.core import Agent, User
from vision_agents.plugins import speechify, gemini, deepgram, getstream

agent = Agent(
    edge=getstream.Edge(),
    agent_user=User(name="Assistant", id="agent"),
    instructions="You are a helpful assistant.",
    llm=gemini.LLM("gemini-3-flash-preview"),
    stt=deepgram.STT(),
    tts=speechify.TTS(),
)

Set SPEECHIFY_API_KEY in your environment or pass api_key directly.

Parameters

Name	Type	Default	Description
`api_key`	`str`	`None`	API key (defaults to `SPEECHIFY_API_KEY` env var)
`voice_id`	`str`	`"geffen_32"`	Voice ID to use for TTS responses. See the Speechify docs for options
`model`	`str`	`"simba-3.2"`	Speechify TTS model (see the docs for more models)
`language`	`str`	`None`	Optional ISO 639-1 + ISO 3166-1 language tag (e.g. `en-US`). Auto-detected when omitted

Next Steps

Build a Voice Agent

Get started with voice

Build a Video Agent

Add video processing

Sarvam TTS xAI TTS

⌘I

Overview

Language Models

Realtime

Speech-to-Text

Text-to-Speech

Vision & Video

Avatars

Turn Detection

Infrastructure

Edge Transport

Telephony

Custom Integrations

Installation

Quick Start

Parameters

Next Steps

Build a Voice Agent

Build a Video Agent

​Installation

​Quick Start

​Parameters

​Next Steps

Build a Voice Agent

Build a Video Agent

Installation

Quick Start

Parameters

Next Steps