feat: add MiniMax TTS support to talk-llama example

[octo-patch]

Summary

This PR adds MiniMax as a TTS provider option to the talk-llama example, alongside the existing ElevenLabs integration.

Changes

• examples/talk-llama/minimax-tts.py — New standalone Python TTS client that:

  • Calls the MiniMax t2a_v2 API with SSE streaming
  • Decodes hex-encoded audio chunks into an MP3 file
  • Supports 6 English voices (see --list)
  • Matches the CLI interface of eleven-labs.py (-v, -n, -s, -p flags)
  • Requires only the Python standard library (no extra pip install)

• examples/talk-llama/speak — Updated to detect MINIMAX_API_KEY and use MiniMax TTS as a fallback when ElevenLabs is unavailable.

Usage

# Set your MiniMax API key
export MINIMAX_API_KEY=your_api_key

# List available voices
python3 examples/talk-llama/minimax-tts.py --list

# Save TTS to file
python3 examples/talk-llama/minimax-tts.py -s output.mp3 input.txt

# Play directly (requires ffplay)
python3 examples/talk-llama/minimax-tts.py -p input.txt

# Or let the speak script auto-detect it
./examples/talk-llama/speak 0 input.txt

API Reference

• TTS: https://platform.minimax.io/docs/api-reference/speech-t2a-http