TinyAGI · a692570 · Feb 13, 2026
diff --git a/.env.example b/.env.example
@@ -0,0 +1,21 @@
+# TinyClaw Environment Variables
+# Copy this file to .env and fill in your values
+
+# Discord Bot Token
+# Get one at: https://discord.com/developers/applications
+DISCORD_BOT_TOKEN=your_discord_bot_token_here
+
+# Telegram Bot Token
+# Create a bot via @BotFather on Telegram
+TELEGRAM_BOT_TOKEN=your_telegram_bot_token_here
+
+# Telnyx Voice Channel (Optional)
+# Get your API key at: https://portal.telnyx.com/#/app/api-keys
+TELNYX_API_KEY=your_telnyx_api_key_here
+TELNYX_PUBLIC_KEY=your_telnyx_public_key_here
+TELNYX_CONNECTION_ID=your_connection_id_here
+TELNYX_PHONE_NUMBER=+15551234567
+
+# Voice Webhook Configuration
+TELNYX_WEBHOOK_PORT=8080
+TELNYX_WEBHOOK_PATH=/telnyx-webhook
diff --git a/README.md b/README.md
@@ -9,7 +9,7 @@ Run multiple AI agents simultaneously with isolated workspaces and conversation
 - ✅ **Multi-agent** - Run multiple isolated AI agents with specialized roles
 - ✅ **Team collaboration** - Agents hand off work to teammates via chain execution and fan-out
 - ✅ **Multiple AI providers** - Anthropic Claude (Sonnet/Opus) and OpenAI (GPT/Codex)
-- ✅ **Multi-channel** - Discord, WhatsApp, and Telegram
+- ✅ **Multi-channel** - Discord, WhatsApp, Telegram, and Voice (Telnyx/ClawdTalk)
 - ✅ **Parallel processing** - Agents process messages concurrently
 - ✅ **Live TUI dashboard** - Real-time team visualizer for monitoring agent chains
 - ✅ **Persistent sessions** - Conversation context maintained across restarts
@@ -97,6 +97,16 @@ After starting TinyClaw, scan the QR code:
 📱 Settings → Linked Devices → Link a Device
 ```
 
+### Voice Setup (Telnyx/ClawdTalk)
+
+1. Create a [Telnyx account](https://telnyx.com)
+2. Create an API key at [Portal > API Keys](https://portal.telnyx.com/#/app/api-keys)
+3. Configure a [Voice Profile](https://portal.telnyx.com/#/app/voice/profiles)
+4. Purchase a [Phone Number](https://portal.telnyx.com/#/app/numbers)
+5. Set webhook URL in your voice profile
+
+See [docs/VOICE.md](docs/VOICE.md) for detailed setup instructions.
+
 </details>
 
 ## 📋 Commands
@@ -187,7 +197,7 @@ export TINYCLAW_SKIP_UPDATE_CHECK=1
 
 ### In-Chat Commands
 
-These commands work in Discord, Telegram, and WhatsApp:
+These commands work in Discord, Telegram, WhatsApp, and Voice calls:
 
 | Command             | Description                                  | Example                              |
 | ------------------- | -------------------------------------------- | ------------------------------------ |
@@ -266,7 +276,7 @@ See [docs/AGENTS.md](docs/AGENTS.md) for:
 ```
 ┌─────────────────────────────────────────────────────────────┐
 │                     Message Channels                         │
-│         (Discord, Telegram, WhatsApp, Heartbeat)            │
+│    (Discord, Telegram, WhatsApp, Voice/Telnyx, Heartbeat)   │
 └────────────────────┬────────────────────────────────────────┘
                      │ Write message.json
                      ↓
@@ -454,6 +464,7 @@ All channels share agent conversations!
 - [AGENTS.md](docs/AGENTS.md) - Agent management and routing
 - [TEAMS.md](docs/TEAMS.md) - Team collaboration, chain execution, and visualizer
 - [QUEUE.md](docs/QUEUE.md) - Queue system and message flow
+- [VOICE.md](docs/VOICE.md) - Voice channel setup (Telnyx/ClawdTalk)
 - [TROUBLESHOOTING.md](docs/TROUBLESHOOTING.md) - Common issues and solutions
 
 ## 🐛 Troubleshooting
@@ -493,6 +504,7 @@ tinyclaw logs all
 - Inspired by [OpenClaw](https://openclaw.ai/) by Peter Steinberger
 - Built on [Claude Code](https://claude.com/claude-code) and [Codex CLI](https://docs.openai.com/codex)
 - Uses [discord.js](https://discord.js.org/), [whatsapp-web.js](https://github.com/pedroslopez/whatsapp-web.js), [node-telegram-bot-api](https://github.com/yagop/node-telegram-bot-api)
+- Voice powered by [Telnyx](https://telnyx.com) and [ClawdTalk](https://clawdtalk.com)
 
 ## 📄 License
 

diff --git a/docs/VOICE.md b/docs/VOICE.md
@@ -0,0 +1,203 @@
+# Voice Channel (Telnyx/ClawdTalk)
+
+TinyClaw supports voice calling capabilities through Telnyx, powered by ClawdTalk for AI voice interactions. This channel enables your agents to make and receive phone calls.
+
+## Features
+
+- **Inbound calls**: Receive calls on your Telnyx phone number
+- **Outbound calls**: Make calls programmatically from agents
+- **Text-to-speech**: Speak responses to callers
+- **DTMF gathering**: Collect keypad input from callers
+- **Call recording**: Optional recording of conversations
+- **Multi-agent routing**: Route calls to specific agents with `@agent_id`
+
+## Setup
+
+### 1. Create a Telnyx Account
+
+1. Sign up at [telnyx.com](https://telnyx.com)
+2. Navigate to the [Portal](https://portal.telnyx.com)
+3. Complete account verification
+
+### 2. Create an API Key
+
+1. Go to [API Keys](https://portal.telnyx.com/#/app/api-keys)
+2. Click "Create API Key"
+3. Save the key securely (you will need it for configuration)
+
+### 3. Configure a Voice Profile
+
+1. Go to [Voice > Profiles](https://portal.telnyx.com/#/app/voice/profiles)
+2. Create a new voice profile
+3. Note the **Connection ID** for configuration
+
+### 4. Purchase a Phone Number
+
+1. Go to [Phone Numbers](https://portal.telnyx.com/#/app/numbers)
+2. Search for and purchase a number
+3. Assign it to your voice profile
+4. Note the number in E.164 format (e.g., `+15551234567`)
+
+### 5. Configure Webhook
+
+1. In your voice profile settings, set the webhook URL:
+   ```
+   https://your-server.com:8080/telnyx-webhook
+   ```
+2. Ensure your server is accessible from the internet
+3. Port 8080 is used by default (configurable via `TELNYX_WEBHOOK_PORT`)
+
+### 6. Run Setup Wizard
+
+```bash
+tinyclaw setup
+```
+
+When prompted:
+1. Select "Voice (Telnyx/ClawdTalk)" as a channel
+2. Enter your Telnyx API key
+3. Enter your Connection ID
+4. Enter your phone number
+
+## Environment Variables
+
+The voice channel can also be configured via environment variables:
+
+| Variable | Description | Required |
+|----------|-------------|----------|
+| `TELNYX_API_KEY` | Your Telnyx API key | Yes |
+| `TELNYX_PUBLIC_KEY` | Your Telnyx public key (for webhook verification) | Recommended |
+| `TELNYX_CONNECTION_ID` | Voice profile connection ID | For outbound calls |
+| `TELNYX_PHONE_NUMBER` | Your Telnyx phone number (E.164) | For outbound calls |
+| `TELNYX_WEBHOOK_PORT` | Webhook server port (default: 8080) | No |
+| `TELNYX_WEBHOOK_PATH` | Webhook endpoint path (default: `/telnyx-webhook`) | No |
+
+## Usage
+
+### Inbound Calls
+
+When someone calls your Telnyx number:
+1. The call is automatically answered
+2. A greeting is spoken to the caller
+3. The call is routed to the default agent
+4. Agent responses are spoken to the caller
+5. DTMF input is collected and processed
+
+### Outbound Calls
+
+From the command line:
+```bash
+node dist/channels/telnyx-voice-client.js call +15551234567 "Hello, this is a test call"
+```
+
+From within an agent response:
+```
+[voice_call: +15551234567]
+```
+
+### Agent Routing
+
+Route voice calls to specific agents:
+- During a call, say `@agent_id` followed by your request
+- Example: "Let me connect you to @coder for that technical question"
+
+### DTMF Input
+
+Callers can use their keypad to provide input:
+- Press 1 for sales
+- Press 2 for support
+- etc.
+
+The DTMF tones are captured and forwarded to the agent as text.
+
+## Response Metadata
+
+Agents can include special metadata in their responses to control the call:
+
+```json
+{
+  "message": "Thank you for calling. Goodbye!",
+  "metadata": {
+    "speak": "Thank you for calling. Goodbye!",
+    "hangup": true
+  }
+}
+```
+
+| Field | Description |
+|-------|-------------|
+| `speak` | Text to speak to the caller |
+| `hangup` | Whether to end the call after speaking |
+| `gather` | Whether to collect DTMF input after speaking |
+
+## ClawdTalk Integration
+
+[ClawdTalk](https://clawdtalk.com) provides advanced voice AI capabilities for AI agents:
+
+- Natural language understanding
+- Real-time transcription
+- Conversation flow management
+- Multi-language support
+
+To use ClawdTalk with your TinyClaw voice channel:
+1. Sign up at [clawdtalk.com](https://clawdtalk.com)
+2. Follow the integration guide at [github.com/team-telnyx/clawdtalk-client](https://github.com/team-telnyx/clawdtalk-client)
+3. Configure your agent to use ClawdTalk for voice processing
+
+## Troubleshooting
+
+### Webhook Not Receiving Events
+
+1. Verify your webhook URL is publicly accessible
+2. Check firewall rules allow inbound connections on the webhook port
+3. Verify the webhook URL matches what's configured in Telnyx portal
+
+### Outbound Calls Failing
+
+1. Verify `TELNYX_CONNECTION_ID` and `TELNYX_PHONE_NUMBER` are set
+2. Check your Telnyx account has sufficient balance
+3. Verify the destination number is in E.164 format
+
+### Signature Verification Failing
+
+1. Ensure `TELNYX_PUBLIC_KEY` is set correctly
+2. Verify you're using the correct public key (not the API key)
+
+## Logs
+
+View voice channel logs:
+```bash
+tinyclaw logs voice
+```
+
+Or directly:
+```bash
+tail -f ~/.tinyclaw/logs/voice.log
+```
+
+## API Reference
+
+The voice channel exports the following functions for programmatic use:
+
+```typescript
+import { makeOutboundCall, speakOnCall, hangupCall, gatherInput } from './channels/telnyx-voice-client';
+
+// Make an outbound call
+const callControlId = await makeOutboundCall('+15551234567', 'Hello from AI');
+
+// Speak on an active call
+await speakOnCall(callControlId, 'This is a test message');
+
+// Gather DTMF input
+await gatherInput(callControlId);
+
+// Hang up a call
+await hangupCall(callControlId);
+```
+
+## Resources
+
+- [Telnyx Documentation](https://developers.telnyx.com)
+- [Telnyx Node SDK](https://www.npmjs.com/package/@telnyx/node)
+- [ClawdTalk](https://clawdtalk.com)
+- [ClawdTalk Client](https://github.com/team-telnyx/clawdtalk-client)
diff --git a/lib/common.sh b/lib/common.sh
@@ -31,30 +31,35 @@ NC='\033[0m'
 # --- Channel registry ---
 # Single source of truth. Add new channels here and everything else adapts.
 
-ALL_CHANNELS=(discord whatsapp telegram)
+ALL_CHANNELS=(discord whatsapp telegram voice)
 
 declare -A CHANNEL_DISPLAY=(
     [discord]="Discord"
     [whatsapp]="WhatsApp"
     [telegram]="Telegram"
+    [voice]="Voice (Telnyx)"
 )
 declare -A CHANNEL_SCRIPT=(
     [discord]="dist/channels/discord-client.js"
     [whatsapp]="dist/channels/whatsapp-client.js"
     [telegram]="dist/channels/telegram-client.js"
+    [voice]="dist/channels/telnyx-voice-client.js"
 )
 declare -A CHANNEL_ALIAS=(
     [discord]="dc"
     [whatsapp]="wa"
     [telegram]="tg"
+    [voice]="vc"
 )
 declare -A CHANNEL_TOKEN_KEY=(
     [discord]="discord_bot_token"
     [telegram]="telegram_bot_token"
+    [voice]="api_key"
 )
 declare -A CHANNEL_TOKEN_ENV=(
     [discord]="DISCORD_BOT_TOKEN"
     [telegram]="TELEGRAM_BOT_TOKEN"
+    [voice]="TELNYX_API_KEY"
 )
 
 # Runtime state: filled by load_settings
@@ -105,7 +110,11 @@ load_settings() {
     for ch in "${ALL_CHANNELS[@]}"; do
         local token_key="${CHANNEL_TOKEN_KEY[$ch]:-}"
         if [ -n "$token_key" ]; then
-            CHANNEL_TOKENS[$ch]=$(jq -r ".channels.${ch}.bot_token // empty" "$SETTINGS_FILE" 2>/dev/null)
+            if [ "$ch" = "voice" ]; then
+                CHANNEL_TOKENS[$ch]=$(jq -r ".channels.voice.api_key // empty" "$SETTINGS_FILE" 2>/dev/null)
+            else
+                CHANNEL_TOKENS[$ch]=$(jq -r ".channels.${ch}.bot_token // empty" "$SETTINGS_FILE" 2>/dev/null)
+            fi
         fi
     done
 

diff --git a/lib/setup-wizard.sh b/lib/setup-wizard.sh
@@ -19,24 +19,28 @@ echo ""
 
 # --- Channel registry ---
 # To add a new channel, add its ID here and fill in the config arrays below.
-ALL_CHANNELS=(telegram discord whatsapp)
+ALL_CHANNELS=(telegram discord whatsapp voice)
 
 declare -A CHANNEL_DISPLAY=(
     [telegram]="Telegram"
     [discord]="Discord"
     [whatsapp]="WhatsApp"
+    [voice]="Voice (Telnyx/ClawdTalk)"
 )
 declare -A CHANNEL_TOKEN_KEY=(
     [discord]="discord_bot_token"
     [telegram]="telegram_bot_token"
+    [voice]="telnyx_api_key"
 )
 declare -A CHANNEL_TOKEN_PROMPT=(
     [discord]="Enter your Discord bot token:"
     [telegram]="Enter your Telegram bot token:"
+    [voice]="Enter your Telnyx API key:"
 )
 declare -A CHANNEL_TOKEN_HELP=(
     [discord]="(Get one at: https://discord.com/developers/applications)"
     [telegram]="(Create a bot via @BotFather on Telegram to get a token)"
+    [voice]="(Get one at: https://portal.telnyx.com/#/app/api-keys)"
 )
 
 # Channel selection - simple checklist
@@ -265,6 +269,18 @@ CHANNELS_JSON="${CHANNELS_JSON}]"
 # Build channel configs with tokens
 DISCORD_TOKEN="${TOKENS[discord]:-}"
 TELEGRAM_TOKEN="${TOKENS[telegram]:-}"
+VOICE_API_KEY="${TOKENS[voice]:-}"
+
+# Collect additional voice configuration if voice is enabled
+VOICE_CONNECTION_ID=""
+VOICE_PHONE_NUMBER=""
+if [[ " ${ENABLED_CHANNELS[*]} " =~ " voice " ]] && [ -n "$VOICE_API_KEY" ]; then
+    echo ""
+    echo -e "${YELLOW}Voice channel configuration:${NC}"
+    echo ""
+    read -rp "  Telnyx Connection ID: " VOICE_CONNECTION_ID
+    read -rp "  Telnyx Phone Number (E.164 format, e.g., +15551234567): " VOICE_PHONE_NUMBER
+fi
 
 # Write settings.json with layered structure
 # Use jq to build valid JSON to avoid escaping issues with agent prompts
@@ -288,7 +304,12 @@ cat > "$SETTINGS_FILE" <<EOF
     "telegram": {
       "bot_token": "${TELEGRAM_TOKEN}"
     },
-    "whatsapp": {}
+    "whatsapp": {},
+    "voice": {
+      "api_key": "${VOICE_API_KEY}",
+      "connection_id": "${VOICE_CONNECTION_ID}",
+      "phone_number": "${VOICE_PHONE_NUMBER}"
+    }
   },
   ${AGENTS_JSON}
   ${MODELS_SECTION},