Claude Talk Setup

Run the complete setup for Claude Talk TTS. This will install dependencies, download models, let the user choose a voice, and configure the hooks.

Step 1: Install Dependencies

cd /Users/pv/git/claude-talk && uv sync

Step 2: Download Models

Download the Kokoro ONNX models (~340MB total):

mkdir -p /Users/pv/git/claude-talk/models
curl -L -o /Users/pv/git/claude-talk/models/kokoro-v1.0.onnx https://github.com/thewh1teagle/kokoro-onnx/releases/download/model-files-v1.0/kokoro-v1.0.onnx
curl -L -o /Users/pv/git/claude-talk/models/voices-v1.0.bin https://github.com/thewh1teagle/kokoro-onnx/releases/download/model-files-v1.0/voices-v1.0.bin

Verify the models downloaded:

ls -la /Users/pv/git/claude-talk/models/

Step 3: Voice Selection

Play voice samples for the user to choose from. For each voice, run the command and let them hear it:

Voice 1 - Heart (Female, Warm):

uv run --directory /Users/pv/git/claude-talk python -c "
import sys
sys.path.insert(0, 'src')
from claude_talk.tts import KokoroTTS
tts = KokoroTTS()
tts.speak('Why do programmers prefer dark mode? Because light attracts bugs.', voice='af_heart')
"

Voice 2 - Michael (Male, Professional):

uv run --directory /Users/pv/git/claude-talk python -c "
import sys
sys.path.insert(0, 'src')
from claude_talk.tts import KokoroTTS
tts = KokoroTTS()
tts.speak('Why do Java developers wear glasses? Because they cant C sharp.', voice='am_michael')
"

Voice 3 - Bella (Female, Expressive):

uv run --directory /Users/pv/git/claude-talk python -c "
import sys
sys.path.insert(0, 'src')
from claude_talk.tts import KokoroTTS
tts = KokoroTTS()
tts.speak('There are only 10 kinds of people. Those who understand binary, and those who dont.', voice='af_bella')
"

Voice 4 - Fenrir (Male, Deep):

uv run --directory /Users/pv/git/claude-talk python -c "
import sys
sys.path.insert(0, 'src')
from claude_talk.tts import KokoroTTS
tts = KokoroTTS()
tts.speak('A SQL query walks into a bar, sees two tables, and asks... can I join you?', voice='am_fenrir')
"

Ask which voice they prefer after playing the samples.

Step 4: Save Configuration

Once they choose a voice, save the config (replace VOICE_ID with the chosen voice, e.g., 'af_heart'):

uv run --directory /Users/pv/git/claude-talk python -c "
import sys
sys.path.insert(0, 'src')
from claude_talk.config import Config, save_config
config = Config(
    enabled=True,
    voice='VOICE_ID',
    max_chars=500
)
save_config(config)
print('Configuration saved!')
"

Step 5: Install Hook

Install the Claude Code hooks:

uv run --directory /Users/pv/git/claude-talk python -c "
import sys
sys.path.insert(0, 'src')
from claude_talk.setup import install_hook
install_hook()
print('Hook installed!')
"

Done!

Tell the user setup is complete and remind them of the commands:

• /claude-talk:disable - Turn off speech
• /claude-talk:enable - Turn on speech
• /claude-talk:voice - Change voice

Note: They'll need to restart Claude Code for the hook to take effect.