pipecat-ten-vad 0.1.0

TEN VAD integration for Pipecat — voice activity detection for real-time voice agents

pipecat-ten-vad

TEN VAD implementation with pipecat

pipecat-ten-vad

TEN VAD integration for Pipecat — a low-latency, high-performance voice activity detector for real-time voice agents.

Community Integration — This package is maintained by the community, not the Pipecat core team. See the Pipecat community integrations page for more.

---

Why TEN VAD?

Most Pipecat pipelines use Silero VAD by default. TEN VAD is worth considering if you need:

• Lower latency — detects speech-to-silence transitions faster, reducing end-of-turn delay in voice agents
• Better short-silence handling — correctly identifies brief pauses between words that Silero VAD can miss
• Lower compute — smaller model with a lower real-time factor (RTF) across tested platforms

---

Pipecat Compatibility

Tested with Pipecat v0.0.106.

Check the Pipecat changelog when upgrading.

---

Installation

Step 1 — Install TEN VAD

TEN VAD is not yet available on PyPI. Install it directly from GitHub:

pip install git+https://github.com/TEN-framework/ten-vad.git

Step 2 — Install pipecat-ten-vad

pip install pipecat-ten-vad

Step 3 — Install Pipecat with your required extras

pip install "pipecat-ai[daily,deepgram,openai,cartesia]"

---

Quick Start

from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.transports.services.daily import DailyParams, DailyTransport
from pipecat_ten_vad import TenVadAnalyzer

transport = DailyTransport(
    room_url,
    token,
    "My Bot",
    DailyParams(
        audio_in_enabled=True,
        audio_out_enabled=True,
        vad_enabled=True,
        vad_analyzer=TenVadAnalyzer(
            sample_rate=16000,
            threshold=0.6,
            params=VADParams(stop_secs=0.3),
        ),
        vad_audio_passthrough=True,
    ),
)

---

Configuration

TenVadAnalyzer parameters

Parameter	Type	Default	Description
sample_rate	int	None	Audio sample rate in Hz. Must be 16000 if provided.
threshold	float	0.6	Detection threshold (0.0–1.0). Higher = requires stronger voice signal to trigger.
params	VADParams	None	Pipecat VAD smoothing params (e.g. stop_secs).

VADParams — key fields

Parameter	Default	Description
stop_secs	0.3	Seconds of silence before end-of-turn is triggered. Lower = more responsive, higher = fewer false cuts.

---

Running the Example

# Clone the repo
git clone https://github.com/rahulsolanki001/pipecat-ten-vad.git
cd pipecat-ten-vad

# Install dependencies
pip install git+https://github.com/TEN-framework/ten-vad.git
pip install "pipecat-ten-vad[examples]"

# Set up environment
cp .env.example .env
# Edit .env with your API keys

# Run
python examples/foundational/ten_vad_basic.py

---

Requirements

• Python >= 3.10, < 3.13
• pipecat-ai >= 0.0.106
• numpy >= 1.24
• TEN VAD (installed from GitHub — see above)

Supported platforms: Linux x64, macOS (arm64 + x64), Windows x64

---

Contributing

Issues and PRs are welcome. When reporting a bug, please include:

• Your Python version
• Your pipecat-ai version (pip show pipecat-ai)
• A minimal reproduction script

---

License

MIT — see LICENSE for details.

TEN VAD itself is licensed under Apache 2.0 with additional conditions — see the TEN VAD license.

---

Acknowledgements

• Pipecat by Daily
• TEN VAD by the TEN Framework team