Sinapsis CSM

🐍 Installation • 🚀 Features • 📚 Usage example • 📙 Documentation • 🔍 License

This Sinapsis CSM package integrates a lightweight, efficient text-to-speech engine using the CSM model. It provides a simple template to convert input text into speech using Sinapsis.

🐍 Installation

[!IMPORTANT] Sinapsis project requires Python 3.10 or higher.

Install using your preferred package manager. We strongly recommend using uv. To install uv, refer to the official documentation.

Install with uv:

uv pip install sinapsis-csm --extra-index-url https://pypi.sinapsis.tech

Or with raw pip:

pip install sinapsis-csm --extra-index-url https://pypi.sinapsis.tech

[!IMPORTANT] Templates in each package may require additional dependencies. For development, we recommend installing the package with all the optional dependencies:

With uv:

uv pip install sinapsis-csm[all] --extra-index-url https://pypi.sinapsis.tech

Or with raw pip:

pip install sinapsis-csm[all] --extra-index-url https://pypi.sinapsis.tech

To run this package you need a HuggingFace token. See the official instructions and set it using

export HF_TOKEN=<token-provided-by-hf>

and test it through the cli or the webapp.

Access to the following models is needed:

• Llama-3.2-1B
• CSM-1B

🚀 Features

Templates Supported

• CSMTTS: Converts text into speech using the CSM model.

  Attributes

  • speaker_id (int, default: 0): Speaker identity index.
  • max_audio_length_ms (int, default: 10000): Max audio length in milliseconds.
  • device ("cpu" or "cuda", default: "cpu"): Device used for inference.
  • context (context: list[str] | None = None): Optional list of past utterances for context.
  • sample_rate_hz (int, default: 24000): Output audio sample rate.

📚 Usage example

This example shows how to use the CSMTTS template to convert text into speech and save it to disk.

agent:
  name: csm_tts_agent
  description: Agent that synthesizes speech from text using the CSM model.

templates:
  - template_name: InputTemplate
    class_name: InputTemplate
    attributes: {}

  - template_name: TextInput
    class_name: TextInput
    template_input: InputTemplate
    attributes:
      text: "Hi, my name is Taylor and this is Sinapsis"

  - template_name: CSMTTS
    class_name: CSMTTS
    template_input: TextInput
    attributes:
      speaker_id: 0
      max_audio_length_ms: 10000
      device: cpu
      context: null
      sample_rate_hz: 24000

  - template_name: AudioWriterSoundFile
    class_name: AudioWriterSoundFile
    template_input: CSMTTS
    attributes:
      save_dir: csm_tts
      extension: wav

To run the config, use:

sinapsis run packages/sinapsis_csm/src/sinapsis_csm/configs/csm_agent.yml

[!NOTE] The TextInput and AudioWriterSoundFile templates come from the sinapsis-data-readers and sinapsis-data-writers packages. Make sure they are installed to use this example.

🌐 Webapp

[!IMPORTANT] To run the app you first need to clone this repository:

git clone git@github.com:Sinapsis-ai/sinapsis-speech.git
cd sinapsis-speech

[!NOTE] If you'd like to enable external app sharing in Gradio, export GRADIO_SHARE_APP=True

docker compose -f docker/compose.yaml build

docker compose -f docker/compose_apps.yaml up -d sinapsis-csm

docker logs -f sinapsis-csm

Running on local URL:  http://127.0.0.1:7860

docker compose -f docker/compose_apps.yaml down

uv sync --frozen

uv pip install sinapsis-speech[all] --extra-index-url https://pypi.sinapsis.tech

uv run webapps/packet_tts_apps/csm_tts_app.py

Running on local URL:  http://127.0.0.1:7860

📙 Documentation

Documentation is available on the Sinapsis website.

Tutorials and guides for different templates and agents are available at docs.sinapsis.tech/tutorials.

🔍 License

This project is licensed under the AGPLv3 license, which encourages open collaboration and sharing. For more details, please refer to the LICENSE file.

For commercial use, please refer to our official Sinapsis website for information on obtaining a commercial license.