agentvibes 2.0.17-beta.2 → 2.0.17-beta.21

package/INSTALL_MCP_WINDOWS.md

@@ -1,75 +1,153 @@
-# Install MCP Package on Windows
+# Install AgentVibes MCP on Windows
 
-## Issue Found
+## Prerequisites
 
-The MCP server logs show:
+1. **Python 3.10+** - Download from [python.org](https://python.org)
+2. **Node.js 16+** - Should already be installed if you have Claude Desktop
+3. **ElevenLabs API Key** (optional) - Get from [elevenlabs.io](https://elevenlabs.io)
+
+## Installation
+
+### Option 1: NPX (Recommended)
+
+AgentVibes MCP is published to npm and can be installed automatically via NPX.
+
+#### Step 1: Set Environment Variable
+
+Open **PowerShell** and set your ElevenLabs API key (if using ElevenLabs):
+
+```powershell
+setx ELEVENLABS_API_KEY "your-api-key-here"
+```
+
+If you don't have an API key, Piper TTS (free, offline) will be used automatically.
+
+#### Step 2: Configure Claude Desktop
+
+Open your Claude Desktop config file:
+```
+C:\Users\YOUR_USERNAME\AppData\Roaming\Claude\claude_desktop_config.json
 ```
-ModuleNotFoundError: No module named 'mcp'
+
+Add this to the `mcpServers` section:
+
+```json
+{
+  "mcpServers": {
+    "agentvibes": {
+      "command": "npx",
+      "args": [
+        "-y",
+        "-p",
+        "agentvibes@beta",
+        "agentvibes-mcp-server"
+      ],
+      "env": {
+        "ELEVENLABS_API_KEY": "${ELEVENLABS_API_KEY}"
+      }
+    }
+  }
+}
 ```
 
-This means Windows Python doesn't have the `mcp` package installed.
+**⚠️ IMPORTANT:** The `-p` flag is critical for NPX to correctly parse the command!
+
+#### Step 3: Restart Claude Desktop
+
+1. **Completely exit** Claude Desktop (right-click system tray → Exit)
+2. **Wait 5 seconds**
+3. **Restart** Claude Desktop
+
+The first launch will download AgentVibes and install Python dependencies automatically.
+
+#### Step 4: Verify
+
+In Claude Desktop, try:
+- "List available voices"
+- "Use text to speech to say hello"
+
+## Troubleshooting
 
-## Solution
+### Error: "No module named 'mcp'"
 
-Open **PowerShell** or **Command Prompt** as Administrator and run:
+The Python MCP package is missing. Install it:
 
 ```powershell
 pip install mcp
 ```
 
-Or if you have multiple Python versions:
-
+Or:
 ```powershell
 python -m pip install mcp
 ```
 
-## Verify Installation
+### Error: "Python 3 Not Found"
 
-```powershell
-python -c "import mcp; print('MCP installed successfully!')"
+Install Python 3.10+ from [python.org](https://python.org), then restart Claude Desktop.
+
+### Error: "bin/mcp-server: No such file or directory"
+
+You're using the wrong NPX syntax. Make sure your config includes the `-p` flag:
+
+```json
+"args": ["-y", "-p", "agentvibes@beta", "agentvibes-mcp-server"]
 ```
 
-## After Installation
+**NOT:**
+```json
+"args": ["-y", "agentvibes@beta", "agentvibes-mcp-server"]  ❌ Missing -p
+```
 
-1. **Restart Claude Desktop** completely
-2. The AgentVibes MCP server should now work!
+### Check Logs
 
-## Test in Claude Desktop
+View the MCP server logs:
+```
+C:\Users\YOUR_USERNAME\AppData\Roaming\Claude\logs\mcp-server-agentvibes.log
+```
 
-Once restarted, try:
-- "What AgentVibes tools do you have?"
-- "Use text to speech to say hello"
+### Clear NPX Cache
 
-## Configuration Summary
+If you're stuck on an old version:
 
-Your Claude Desktop config is now set to:
-- **Server:** `C:\Users\Paul\AgentVibes\mcp-server\server.py`
-- **Python:** Windows Python (the one in your PATH)
-- **Provider:** Piper TTS (free, offline - no API key needed)
+```powershell
+npx clear-npx-cache
+```
 
-## Troubleshooting
+Then restart Claude Desktop to download fresh.
+
+## Configuration Options
+
+### Using Piper TTS (Free, Offline)
+
+Remove or omit the `ELEVENLABS_API_KEY` environment variable. Piper will be used automatically.
+
+### Using ElevenLabs (Premium)
+
+Set your API key:
+```powershell
+setx ELEVENLABS_API_KEY "your-key"
+```
 
-If it still doesn't work after installing `mcp`:
+### Beta vs Stable
 
-1. **Check Python version:**
-   ```powershell
-   python --version
-   ```
-   Should be Python 3.10 or higher
+- **Beta:** Latest features, may have bugs
+  ```json
+  "args": ["-y", "-p", "agentvibes@beta", "agentvibes-mcp-server"]
+  ```
 
-2. **Check pip:**
-   ```powershell
-   pip --version
-   ```
+- **Stable:** Production-ready
+  ```json
+  "args": ["-y", "-p", "agentvibes@latest", "agentvibes-mcp-server"]
+  ```
 
-3. **Try explicit install:**
-   ```powershell
-   C:\Users\Paul\AppData\Local\Programs\Python\Python312\python.exe -m pip install mcp
-   ```
+## Next Steps
 
-4. **Check logs again:**
-   View: `C:\Users\Paul\AppData\Roaming\Claude\logs\mcp-server-agentvibes.log`
+Once installed, explore these slash commands in Claude Code:
+- `/agent-vibes:list` - List available voices
+- `/agent-vibes:switch <voice>` - Change voice
+- `/agent-vibes:personality <name>` - Set personality style
+- `/agent-vibes:provider` - Manage TTS providers
 
 ---
 
-🎤 **Ready to make Claude Desktop speak!**
+🎤 **Ready to make your AI agents speak!**

package/README.md

@@ -123,6 +123,20 @@ Instead of remembering slash commands like `/agent-vibes:switch Aria`, just say:
   - The MCP server will automatically install Python packages (mcp, pipx, Piper TTS)
   - But Python 3 itself must be installed first by you
 
+**Step 1: Set Your ElevenLabs API Key (Windows)**
+
+Open Command Prompt or PowerShell and run:
+```cmd
+setx ELEVENLABS_API_KEY "your-api-key-here"
+```
+
+**Important:** After running `setx`, you must:
+1. Close Claude Desktop completely (if running)
+2. Open a **new** terminal window (the variable won't be available in current terminal)
+3. Restart Claude Desktop
+
+**Step 2: Configure Claude Desktop**
+
 Add to `~/Library/Application Support/Claude/claude_desktop_config.json` (macOS) or `%APPDATA%\Claude\claude_desktop_config.json` (Windows):
 
 ```json
@@ -130,15 +144,14 @@ Add to `~/Library/Application Support/Claude/claude_desktop_config.json` (macOS)
   "mcpServers": {
     "agentvibes": {
       "command": "npx",
-      "args": ["-y", "agentvibes-mcp-server"],
-      "env": {
-        "ELEVENLABS_API_KEY": "${ELEVENLABS_API_KEY}"
-      }
+      "args": ["-y", "agentvibes@beta", "agentvibes-mcp-server"]
     }
   }
 }
 ```
 
+**Note:** The `env` section is optional. Claude Desktop automatically inherits Windows environment variables, so if you've set `ELEVENLABS_API_KEY` with `setx`, it will be available automatically.
+
 #### For Warp Terminal
 
 Add to `~/.warp/mcp.json`:
@@ -147,7 +160,7 @@ Add to `~/.warp/mcp.json`:
 {
   "agentvibes": {
     "command": "npx",
-    "args": ["-y", "agentvibes-mcp-server"],
+    "args": ["-y", "agentvibes@beta", "agentvibes-mcp-server"],
     "env": {
       "ELEVENLABS_API_KEY": "${ELEVENLABS_API_KEY}"
     }
@@ -164,7 +177,7 @@ Add to `.mcp-minimal.json` in your project:
   "mcpServers": {
     "agentvibes": {
       "command": "npx",
-      "args": ["-y", "agentvibes-mcp-server"],
+      "args": ["-y", "agentvibes@beta", "agentvibes-mcp-server"],
       "env": {
         "ELEVENLABS_API_KEY": "${ELEVENLABS_API_KEY}"
       }

package/RELEASE_NOTES.md

@@ -1,5 +1,102 @@
 # 🎤 AgentVibes Release Notes
 
+## 📦 v2.0.17-beta.12 - Sequential Audio & Target Language Replay (2025-10-16)
+
+### 🎯 AI Summary
+
+This release significantly improves the language learning experience by solving audio playback timing issues and adding essential replay functionality. Users can now enjoy seamless sequential audio playback where translations wait for the main language to finish, and a new replay command allows learners to hear target language audio again without triggering new translations.
+
+### ✨ New Features
+
+#### Sequential Audio Playback
+- **No More Overlapping Audio**: Target language now waits for main language to complete
+- **Lock File Mechanism**: Uses `/tmp/agentvibes-audio.lock` for coordination
+- **Precise Timing**: Detects audio duration with ffprobe for accurate sequencing
+- **Auto-timeout**: 30-second safety timeout prevents stuck locks
+
+#### Target Language Replay Command
+- **New Command**: `/agent-vibes:replay-target` replays last target language audio
+- **Perfect for Learning**: Hear translations again without re-triggering TTS
+- **Automatic Tracking**: Saves target audio path to `.claude/last-target-audio.txt`
+- **Works Seamlessly**: Integrated with language learning mode
+
+### 🐛 Bug Fixes
+
+#### Audio Quality
+- **128kbps Bitrate Preserved**: Fixed ffmpeg downsampling from 128kbps to 64kbps
+- **No More Static**: Eliminated quality degradation during audio padding
+- **Added `-b:a 128k` Parameter**: Explicit bitrate specification in ffmpeg
+
+#### JSON Escaping (ElevenLabs)
+- **Robust API Communication**: Now uses jq for proper JSON escaping
+- **Special Characters Fixed**: Handles quotes, backslashes, and newlines correctly
+- **Prevents API Errors**: No more "Invalid \\escape" errors from ElevenLabs
+
+### 📁 Modified Files
+
+- `.claude/hooks/play-tts-piper.sh` - Added sequential playback + target tracking
+- `.claude/hooks/play-tts-elevenlabs.sh` - Added sequential playback + target tracking + JSON fix
+- `.claude/commands/agent-vibes/replay-target.md` - New slash command
+- `.claude/hooks/replay-target-audio.sh` - Replay implementation
+
+### 🎓 User Impact
+
+#### Language Learners
+✅ **Crystal Clear Audio**: No overlapping English and Spanish
+✅ **Practice Listening**: Replay target language as many times as needed
+✅ **Better Quality**: Full 128kbps audio without static or degradation
+
+#### All Users
+✅ **More Reliable**: Proper JSON escaping prevents API failures
+✅ **Smoother Experience**: Sequential playback feels more natural
+
+### 🚀 How to Use
+
+**Enable Language Learning Mode:**
+```bash
+/agent-vibes:learn spanish
+```
+
+**Replay Last Target Audio:**
+```bash
+/agent-vibes:replay-target
+```
+
+**The Flow:**
+1. Ask a question or make a statement
+2. Hear English response (waits to complete)
+3. Hear Spanish translation (plays after English)
+4. Use `/agent-vibes:replay-target` to hear Spanish again
+
+### 🔧 Technical Details
+
+**Sequential Playback Implementation:**
+- Lock file created before audio starts
+- Audio duration detected with ffprobe
+- Background process waits for duration then releases lock
+- Next audio waits up to 30 seconds for lock release
+
+**Target Language Tracking:**
+- Only tracks when `CURRENT_LANGUAGE != "english"`
+- Path saved to project-local `.claude/last-target-audio.txt`
+- Replay script validates file exists before playback
+
+### 📝 Breaking Changes
+
+None - fully backward compatible with existing installations.
+
+### 🐛 Known Issues
+
+None reported.
+
+---
+
+**Previous Release:** v2.0.17-beta.11
+**Release Date:** October 16, 2025
+**Git Commit:** f4d3899
+
+---
+
 ## 📦 v2.0.16 - Remote Audio Setup Guide & Scripts (2025-10-12)
 
 ### 🤖 AI Summary

package/RELEASE_NOTES_V2.md

@@ -91,6 +91,107 @@ AgentVibes now speaks **30+ languages** with native voice quality!
 
 ---
 
+### 🎓 Language Learning Mode (Beta)
+
+**Learn languages naturally with dual-language TTS!**
+
+AgentVibes now includes a **language learning mode** that helps you learn new languages through context and repetition. Every acknowledgment and completion is spoken TWICE - first in your main language (English), then in your target language.
+
+#### **How It Works:**
+
+1. **Set your target language** - Choose from 30+ supported languages
+2. **Enable learning mode** - Activates dual-language TTS
+3. **Natural repetition** - Hear everything twice in context
+4. **Adjustable speed** - Slow down target language for better comprehension
+
+#### **New Learning Commands:**
+
+```bash
+# Set the language you want to learn
+/agent-vibes:target spanish
+/agent-vibes:target french
+/agent-vibes:target german
+
+# Set voice for target language (auto-selected based on provider)
+/agent-vibes:target-voice Antoni          # ElevenLabs
+/agent-vibes:target-voice es_ES-davefx-medium  # Piper
+
+# Enable/disable learning mode
+/agent-vibes:learn
+
+# Set your main/native language
+/agent-vibes:language english
+
+# Adjust speech speed (Piper only)
+/agent-vibes:set-speed 2x          # 2x slower (great for learning)
+/agent-vibes:set-speed target 2x   # Slow down target language only
+/agent-vibes:set-speed normal      # Reset to normal speed
+```
+
+#### **Example Learning Session:**
+
+```
+User: "hello"
+
+Claude (English): "Hey there! Great to hear from you!"
+🔊 Plays in English with your configured voice
+
+Claude (Spanish): "¡Hola! ¡Qué bueno saber de ti!"
+🔊 Plays in Spanish with target voice (Antoni/es_ES-davefx-medium)
+```
+
+#### **Advanced Features:**
+
+**🎚️ Speech Rate Control (Piper TTS):**
+- Slow down target language for better comprehension
+- Separate speed controls for main and target languages
+- Intuitive syntax: `2x` = 2x slower, `0.5x` = 2x faster
+- Perfect for language learners who need more time to process
+
+**🔄 Mixed Provider Support:**
+- Use **ElevenLabs for English** (premium quality)
+- Use **Piper for Spanish** (free, slower speech)
+- System auto-detects provider from voice name
+- Seamless switching between providers
+
+**🎯 Auto-Voice Selection:**
+- System automatically selects the best voice for your target language
+- Provider-aware: ElevenLabs voices for ElevenLabs, Piper voices for Piper
+- Smart fallback if preferred voice unavailable
+
+**🌍 Supported Target Languages:**
+Spanish, French, German, Italian, Portuguese, Chinese, Japanese, Korean, Russian, Arabic, Hindi, Polish, Dutch, Turkish, Swedish, Danish, Norwegian, Finnish, Czech, Romanian, Ukrainian, Greek, Bulgarian, Croatian, Slovak, and more!
+
+#### **Voice Mappings by Provider:**
+
+**ElevenLabs Voices:**
+- Spanish → Antoni
+- French → Rachel
+- German → Domi
+- Italian → Bella
+- Portuguese → Matilda
+- Chinese, Japanese, Korean → Antoni (multilingual)
+
+**Piper Voices (Free, Offline):**
+- Spanish → es_ES-davefx-medium
+- French → fr_FR-siwis-medium
+- German → de_DE-thorsten-medium
+- Italian → it_IT-riccardo-x_low
+- Portuguese → pt_BR-faber-medium
+- Chinese → zh_CN-huayan-medium
+- Japanese → ja_JP-hikari-medium
+
+#### **Why This Helps Learning:**
+
+1. **Context-based learning** - Hear words/phrases in real situations
+2. **Natural repetition** - Every message twice, reinforcing vocabulary
+3. **Pronunciation practice** - Native-quality voices model correct pronunciation
+4. **Adjustable pace** - Slow down difficult phrases with speed control
+5. **Consistent exposure** - Learn while coding, naturally building vocabulary
+6. **No extra effort** - Learning happens passively as you work
+
+---
+
 ### 🎤 Expanded Voice Library (27+ Voices)
 
 **New multilingual voices added:**
@@ -269,6 +370,8 @@ sarcastic, flirty, pirate, grandpa, dry-humor, angry, robot, zen, professional,
 │   ├── provider-manager.sh      # Provider switching
 │   ├── provider-commands.sh     # Provider CLI
 │   ├── language-manager.sh      # Language system
+│   ├── learn-manager.sh         # Language learning mode
+│   ├── speed-manager.sh         # Speech rate control (Piper)
 │   ├── voice-manager.sh         # Voice switching
 │   ├── personality-manager.sh   # Personality system
 │   ├── sentiment-manager.sh     # Sentiment system
@@ -289,6 +392,12 @@ sarcastic, flirty, pirate, grandpa, dry-humor, angry, robot, zen, professional,
 - `tts-personality.txt` - Active personality
 - `tts-sentiment.txt` - Active sentiment
 - `tts-language.txt` - Selected language
+- `tts-learn-mode.txt` - Learning mode status (ON/OFF)
+- `tts-target-language.txt` - Target language for learning
+- `tts-target-voice.txt` - Voice for target language
+- `tts-main-language.txt` - Main/native language
+- `config/piper-speech-rate.txt` - Main voice speech rate (Piper)
+- `config/piper-target-speech-rate.txt` - Target voice speech rate (Piper)
 
 ---
 
@@ -315,6 +424,15 @@ sarcastic, flirty, pirate, grandpa, dry-humor, angry, robot, zen, professional,
 - `/agent-vibes:set-language list` - Show languages
 - `/agent-vibes:set-language english` - Reset to English
 
+### Language Learning Commands:
+- `/agent-vibes:target <language>` - Set target language to learn
+- `/agent-vibes:target-voice <voice>` - Set voice for target language
+- `/agent-vibes:learn` - Enable/disable learning mode
+- `/agent-vibes:language <language>` - Set main/native language
+- `/agent-vibes:set-speed <speed>` - Set speech rate (Piper only)
+- `/agent-vibes:set-speed target <speed>` - Set target language speed
+- `/agent-vibes:set-speed get` - Show current speed settings
+
 ### Personality Commands:
 - `/agent-vibes:personality <name>` - Set personality
 - `/agent-vibes:personality list` - Show all
@@ -407,6 +525,17 @@ sarcastic, flirty, pirate, grandpa, dry-humor, angry, robot, zen, professional,
 
 ## 🐛 Bug Fixes
 
+### v2.0.17-beta Series (Language Learning Mode):
+- **Fixed ElevenLabs audio static** - Added MP3 codec (`-c:a libmp3lame`) to prevent WAV format issues
+- **Fixed MCP provider switching** - Enhanced non-interactive detection with `CLAUDE_PROJECT_DIR` check
+- **Fixed target voice sync** - Auto-updates target voice when switching providers
+- **Fixed voice/provider mismatches** - Output style now lets system choose voice based on active provider
+- **Fixed learning mode config** - Corrected file names (`tts-learn-mode.txt`) and value checks (`ON`/`OFF`)
+- **Fixed Piper speech rate** - Properly reads numeric values from config files (strips comments)
+- **Fixed interactive prompts in MCP** - Provider switch commands now work seamlessly via slash commands
+- **Fixed Spanish voice download** - Voice models now download automatically with user consent
+
+### v2.0.0 Core Fixes:
 - Fixed symlink support for shared hooks
 - Fixed WSL audio static with silence padding
 - Fixed installer directory detection for npx
@@ -419,15 +548,17 @@ sarcastic, flirty, pirate, grandpa, dry-humor, angry, robot, zen, professional,
 
 ## 📊 Statistics
 
-**v2.0.0 by the numbers:**
+**v2.0.17-beta by the numbers:**
 - 🎤 **200+ voices** across both providers
-- 🌍 **30+ languages** supported
+- 🌍 **30+ languages** supported (learning mode for all)
+- 🎓 **1 language learning mode** with dual-language TTS
 - 🎭 **19 personalities** included
 - 💭 **19 sentiments** available
-- 📝 **15 slash commands** total
-- 🔧 **12 hook scripts** installed
-- 🔌 **2 TTS providers** available
+- 📝 **20+ slash commands** total (incl. learning commands)
+- 🔧 **14 hook scripts** installed (incl. learn-manager, speed-manager)
+- 🔌 **2 TTS providers** available (mixed provider support)
 - 🤖 **10 BMAD agents** with voice mapping
+- 🎚️ **Adjustable speech rate** for Piper TTS (0.5x - 3x)
 
 ---
 

package/agentvibes.org/.claude/commands/agent-vibes/add.md

@@ -0,0 +1,21 @@
+---
+description: Add a new custom ElevenLabs TTS voice
+argument-hint: <voice_name> <voice_id>
+---
+
+Add a new custom ElevenLabs TTS voice to your voice library.
+
+Usage:
+- `/agent-vibes:add "My Custom Voice" abc123xyz456789`
+- `/agent-vibes:add Narrator KTPVrSVAEUSJRClDzBw7`
+
+The voice ID should be a 15-30 character alphanumeric string from your ElevenLabs account.
+
+To find your voice IDs:
+1. Go to https://elevenlabs.io/app/voice-library
+2. Click on a voice
+3. Copy the voice ID from the URL or settings
+
+After adding, you can switch to it with `/agent-vibes:switch "Voice Name"`
+
+!bash .claude/hooks/voice-manager.sh add $ARGUMENTS

package/agentvibes.org/.claude/commands/agent-vibes/agent-vibes.md

@@ -0,0 +1,68 @@
+---
+description: ElevenLabs TTS voice management commands
+---
+
+# 🎤 ElevenLabs Voice Management
+
+Manage your ElevenLabs text-to-speech voices with these commands:
+
+## Available Commands
+
+### `/agent-vibes:list [first|last] [N]`
+List all available voices, with optional filtering
+- `/agent-vibes:list` - Show all voices
+- `/agent-vibes:list first 5` - Show first 5 voices
+- `/agent-vibes:list last 3` - Show last 3 voices
+
+### `/agent-vibes:preview [first|last] [N]`
+Preview voices by playing audio samples
+- `/agent-vibes:preview` - Preview first 3 voices
+- `/agent-vibes:preview 5` - Preview first 5 voices
+- `/agent-vibes:preview last 5` - Preview last 5 voices
+
+### `/agent-vibes:switch <voice_name>`
+Switch to a different default voice
+- `/agent-vibes:switch Northern Terry`
+- `/agent-vibes:switch "Cowboy Bob"`
+
+### `/agent-vibes:get`
+Display the currently selected voice
+
+### `/agent-vibes:add <name> <voice_id>`
+Add a new custom voice from your ElevenLabs account
+- `/agent-vibes:add "My Voice" abc123xyz456`
+
+### `/agent-vibes:replay [N]`
+Replay recently played TTS audio
+- `/agent-vibes:replay` - Replay last audio
+- `/agent-vibes:replay 1` - Replay most recent
+- `/agent-vibes:replay 2` - Replay second-to-last
+- `/agent-vibes:replay 3` - Replay third-to-last
+
+Keeps last 10 audio files in history.
+
+### `/agent-vibes:set-pretext <word>`
+Set a prefix word/phrase for all TTS messages
+- `/agent-vibes:set-pretext AgentVibes` - All TTS starts with "AgentVibes:"
+- `/agent-vibes:set-pretext "Project Alpha"` - Custom phrase
+- `/agent-vibes:set-pretext ""` - Clear pretext
+
+Saved locally in `.claude/config/agentvibes.json`
+
+## Getting Voice IDs
+
+To add your own custom voices:
+1. Go to https://elevenlabs.io/app/voice-library
+2. Select or create a voice
+3. Copy the voice ID (15-30 character alphanumeric string)
+4. Use `/agent-vibes:add` to add it
+
+## Default Voices
+
+The system comes with these Character Voices from ElevenLabs:
+- Northern Terry, Grandpa Spuds Oxley, Ms. Walker
+- Ralf Eisend, Amy, Michael, Jessica Anne Bogart
+- Aria, Lutz Laugh, Dr. Von Fusion, Matthew Schmitz
+- Demon Monster, Cowboy Bob, Drill Sergeant
+
+Enjoy your TTS experience! 🎵

package/agentvibes.org/.claude/commands/agent-vibes/commands.json

@@ -0,0 +1,53 @@
+{
+  "namespace": "agent-vibes",
+  "commands": [
+    {
+      "name": "list",
+      "description": "List all available ElevenLabs voices"
+    },
+    {
+      "name": "preview",
+      "description": "Preview ElevenLabs voices by playing audio samples"
+    },
+    {
+      "name": "switch",
+      "description": "Switch to a different ElevenLabs voice"
+    },
+    {
+      "name": "whoami",
+      "description": "Display currently selected voice"
+    },
+    {
+      "name": "sample",
+      "description": "Play a sample with the current or specified voice"
+    },
+    {
+      "name": "replay",
+      "description": "Replay the last TTS message"
+    },
+    {
+      "name": "personality",
+      "description": "Manage AI personality settings"
+    },
+    {
+      "name": "sentiment",
+      "description": "Set temporary personality sentiment"
+    },
+    {
+      "name": "set-pretext",
+      "description": "Configure pre-TTS message text"
+    },
+    {
+      "name": "set-language",
+      "description": "Set TTS language for multilingual voices"
+    },
+    {
+      "name": "add",
+      "description": "Add a new personality"
+    },
+    {
+      "name": "get",
+      "description": "Get personality details"
+    }
+  ]
+}