deepgram

Deepgram Plugin

Speech-to-Text (STT) and Text-to-Speech (TTS) plugins for Vision Agents using the Deepgram API.

Installation

uv add "vision-agents[deepgram]"
# or directly
uv add vision-agents-plugins-deepgram

Speech-to-Text (STT)

High-quality speech recognition using Deepgram's Flux model with built-in turn detection.

from vision_agents.plugins import deepgram

stt = deepgram.STT(
    model="flux-general-en",  # Default model
    eager_turn_detection=True,  # Enable eager end-of-turn detection
)

STT Docs

• https://developers.deepgram.com/docs/flux/quickstart
• https://github.com/deepgram/deepgram-python-sdk/blob/main/examples/listen/v2/connect/async.py

Text-to-Speech (TTS)

Low-latency text-to-speech using Deepgram's Aura model via WebSocket streaming.

from vision_agents.plugins import deepgram

tts = deepgram.TTS(
    model="aura-2-thalia-en",  # Default voice
    sample_rate=16000,  # Audio sample rate
)

Available Voices

Deepgram offers various Aura voice models:

• aura-2-thalia-en - Default female voice
• aura-2-orion-en - Male voice
• See TTS Models for all options

TTS Docs

• https://developers.deepgram.com/docs/tts-websocket
• https://developers.deepgram.com/docs/streaming-text-to-speech

Environment Variables

Set DEEPGRAM_API_KEY in your environment or pass api_key to the constructor.

Example

See the example directory for a complete working example using both STT and TTS.