SeasonalFall84's picture
Merge exp/tts-modal-env: Add TTS AudioRefiner with LLM polish, Modal deployment, and async fixes
8e9e85e

A newer version of the Gradio SDK is available: 6.1.0

Upgrade

Deployments

This directory contains infrastructure deployment scripts for DeepCritical services.

Modal Deployments

TTS Service (modal_tts.py)

Deploys the Kokoro TTS (Text-to-Speech) function to Modal's GPU infrastructure.

Deploy:

modal deploy deployments/modal_tts.py

Features:

  • Kokoro 82M TTS model
  • GPU-accelerated (T4)
  • Voice options: af_heart, af_bella, am_michael, etc.
  • Configurable speech speed

Requirements:

  • Modal account and credentials (MODAL_TOKEN_ID, MODAL_TOKEN_SECRET in .env)
  • GPU quota on Modal

After Deployment: The function will be available at:

  • App: deepcritical-tts
  • Function: kokoro_tts_function

The main application (src/services/tts_modal.py) will call this deployed function.


Adding New Deployments

When adding new deployment scripts:

  1. Create a new file: deployments/<service_name>.py
  2. Use Modal's app pattern:
    import modal
    app = modal.App("deepcritical-<service-name>")
    
  3. Document in this README
  4. Test deployment: modal deploy deployments/<service_name>.py