DEDUPEDWho it is forA game developer voicing characters who need different moods.
The problemFlat TTS delivery breaks immersion.
The outcomeLines delivered with the emotion and emphasis you mark up.
What you need first
- Python with pip
- A CPU is enough for the lightweight models; Kitten-TTS 2 runs realtime on CPUs and consumer GPUs per the README
Steps
- Install kittenml and load Kitten-TTS 2
- Write lines with tags such as [joyful] and emphasis like (((hardly)))
- Generate with preset="expressive" and a voice such as Kiki
- Compare with the default preset for stability
Prompt for your agent
Generate three versions of <line> with KittenTTS using different expression tags and preset "expressive", save each to a WAV, and describe how they differ.
Caveat. The README notes the preset trades stability against expressiveness.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for KittenML/KittenTTS and "Control emotion and emphasis in generated speech"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer