Telex

Your phone is now a Claude Code terminal — via Telegram. Send a message to your Saved Messages, Claude gets it. Claude responds, you see it on Telegram. Text, images, voice notes — in both directions.

Dictate a voice note while driving and Claude starts coding. Send an image from your phone and Claude interprets it. Get spoken responses back in any of 28 voices — all synthesized locally, nothing leaves your machine. Take complete control of your computer using /t to open and manage any number of terminal sessions. Manage multiple Claude sessions from your couch with /s, switch between them, or /kill a stuck one and restart it fresh. Take screenshots of any session using /ss and have the screenshot sent to you via Telegram.

One command to set up. Zero cloud dependencies for voice. Works with any Telegram account.

How it works

Telex has two components:

Watcher daemon — a long-running process that owns the Telegram connection (via the GramJS MTProto library). It delivers incoming messages to iTerm2 by typing them into your Claude session via AppleScript. It also serves a Unix Domain Socket so MCP server instances can send and receive messages without holding their own connection.

MCP server — a thin IPC proxy started by Claude Code. It has no direct Telegram connection. Every tool call is forwarded to the watcher over the socket and the response returned to Claude.

Your phone
    |
    | Telegram (GramJS MTProto)
    |
  Watcher daemon  <── launchd, auto-restarts
    |
    |── AppleScript ──> iTerm2 ──> Claude Code (types message into terminal)
    |
    |── Unix Domain Socket (/tmp/telex-watcher.sock)
              |
              └──> MCP Server (started by Claude Code)
                       |
                       └──> telegram_send / receive / status / wait / login
                            telegram_tts / telegram_speak / telegram_voice_config

The separation means you can have multiple Claude Code sessions open simultaneously. Each MCP server instance registers its TERM_SESSION_ID with the watcher. The first session to register becomes the active recipient for incoming messages; use /N from Telegram or telegram_switch to change the active session.

Architecture

Telex is built on aibroker — a shared core library that provides the platform-independent infrastructure:

aibroker (shared core — installed automatically via npm)
├── Logging, session state, message queuing
├── File persistence (parameterized data directory)
├── Unix Domain Socket IPC (server + client)
├── macOS iTerm2 adapter (AppleScript, tab management)
├── Kokoro TTS (local speech synthesis + Whisper transcription)
└── SessionBackend (deliver messages by typing into iTerm2)

Whazaa (WhatsApp bridge)          Telex (this package)
├── Baileys connection             ├── GramJS MTProto connection
├── WhatsApp formatting            ├── Telegram formatting
├── MCP tools (whatsapp_*)         ├── MCP tools (telegram_*)
└── Desktop DB integration         └── ...

If you install multiple bridges (Whazaa + Telex), each has its own node_modules/ — no conflicts. They share the same iTerm2 session infrastructure at runtime through macOS AppleScript.

Quick start

Tell Claude Code:

Clone https://github.com/mnott/Telex and set it up for me

Claude clones the repo, finds the setup skill, and handles everything autonomously — prerequisites, build, MCP config, watcher daemon, and Telegram authentication. The only thing you do is enter your phone number and verification code when prompted.

Alternative: npx

If you prefer a traditional install without cloning:

npx -y @tekmidian/telex setup

This will:

1. Add Telex to ~/.claude.json
2. Print instructions for starting the watcher and authenticating
3. Credentials are saved to ~/.telex/auth/ on first telex watch

Restart Claude Code. Telex connects automatically from now on.

Manual install

git clone https://github.com/mnott/Telex.git
cd Telex
npm install
npm run build

Then tell Claude to set it up from the local clone.

Prerequisites

• Node.js >= 18
• macOS with iTerm2 for the watch command and iTerm2 delivery
• ffmpeg for TTS voice note conversion (WAV to OGG Opus)
• Whisper for voice note transcription (optional — only needed to receive audio/voice messages)
• Telegram API credentials (TELEGRAM_API_ID and TELEGRAM_API_HASH) from my.telegram.org

Install ffmpeg and Whisper via Homebrew:

brew install ffmpeg
pip install openai-whisper

Getting Telegram API credentials

1. Go to my.telegram.org and log in with your phone number
2. Click "API development tools"
3. Create a new application (any name/description)
4. Copy the api_id (number) and api_hash (string)
5. Set them as environment variables:

export TELEGRAM_API_ID=12345678
export TELEGRAM_API_HASH=abcdef1234567890abcdef1234567890

Add these to your ~/.zshrc or ~/.bashrc for persistence.

How to Use

Once Telex is set up, you talk to Claude in plain language. You never need to know about tool names or parameters — just say what you want.

Sending Messages

Tell Claude what to say and to whom:

• "Send Randolf a Telegram message saying I'll be late"
• "Tell Nicole on Telegram the meeting is moved to 3pm"
• "Message my Saved Messages: pick up milk"

If you don't say who to send it to, Claude sends to your Telegram Saved Messages — useful for notes to yourself.

Voice Notes

Claude can send a Telegram voice note instead of a text message:

• "Send me a voice note saying good morning"
• "Send a voice note to Nicole saying I'm on my way"

You can choose whose voice to use:

• "Say it as George" or "Use George's voice"
• "Use Nicole's voice for this"

See the full voice list in the telegram_tts section below.

Listening Locally (Mac Speakers)

Claude can speak out loud through your Mac — no Telegram needed:

• "Say that out loud"
• "Read that to me"

Voice Mode — Hands-Free

Put Claude into persistent voice mode so every response comes back as audio automatically.

• "Voice mode on" — every Claude response becomes a Telegram voice note
• "Back to text" — back to normal text messages
• "Talk to me locally" — every response plays through your speakers

Voice mode is perfect for driving, cooking, or any time you can't look at a screen.

Screenshots

Send /ss from your phone and the watcher captures the active Claude session's iTerm2 window and sends it back to Telegram as an image.

Session Management (from Your Phone)

Control your Claude sessions from Telegram:

• /s — see a list of your active Claude sessions
• /2 — switch to session 2
• /2 Cooking Project — switch to session 2 and name it
• /n — create a new Claude session

MCP tools

Once configured, Claude Code has 19 tools available:

| Tool | Description | |------|-------------| | telegram_status | Check Telegram connection status | | telegram_send | Send a text message (to Saved Messages or any contact) | | telegram_tts | Send a voice note via TTS (Kokoro, local synthesis) | | telegram_send_file | Send a file (with optional caption and prettify) | | telegram_receive | Drain buffered incoming messages | | telegram_wait | Long-poll for incoming messages (up to timeout) | | telegram_login | Trigger fresh Telegram authentication | | telegram_contacts | List Telegram contacts | | telegram_chats | List Telegram chats | | telegram_history | Get message history for a chat | | telegram_voice_config | Get or set voice/TTS configuration | | telegram_speak | Play TTS locally through Mac speakers | | telegram_rename | Rename the current Claude session | | telegram_restart | Restart the current Claude session | | telegram_discover | Discover and prune Claude sessions | | telegram_sessions | List all Claude sessions | | telegram_switch | Switch to a different Claude session | | telegram_end_session | End (close) a Claude session | | telegram_command | Execute a slash command directly (bypass Telegram round-trip) |

telegram_status

Takes no parameters. Returns connection state, authenticated phone number, and watcher uptime.

telegram_send

Sends a text message to your Saved Messages or any contact.

Parameters:

• message (required) — the text to send
• recipient (optional) — username, phone number, chat ID, or display name; omit for Saved Messages
• voice (optional, boolean) — if true, send as a TTS voice note instead of text

telegram_tts

Converts text to speech and sends it as a Telegram voice note.

• Uses Kokoro-js — 100% local, no internet required after first run
• The model (~160 MB) is downloaded on first use and cached locally
• Requires ffmpeg for WAV to OGG Opus conversion

Parameters:

• text (required) — text to convert to speech
• recipient (optional) — username, phone number, chat ID, or display name; omit for Saved Messages
• voice (optional) — voice name from the table below; omit to use the configured default

Available voices (28 total):

| Category | Voices | |----------|--------| | American Female | af_heart, af_alloy, af_aoede, af_bella, af_jessica, af_kore, af_nicole, af_nova, af_river, af_sarah, af_sky | | American Male | am_adam, am_echo, am_eric, am_fenrir, am_liam, am_michael, am_onyx, am_puck, am_santa | | British Female | bf_alice, bf_emma, bf_isabella, bf_lily | | British Male | bm_daniel, bm_fable, bm_george, bm_lewis |

Default voice: bm_fable

telegram_send_file

Sends a file as a Telegram document, or optionally as inline formatted messages.

Parameters:

• filePath (required) — absolute path to the file
• recipient (optional) — username, phone number, chat ID, or display name; omit for Saved Messages
• caption (optional) — caption text to attach to the file
• prettify (optional, boolean) — if true, send the file contents as formatted inline Telegram messages rather than a document attachment; useful for markdown files you want to read in chat

telegram_receive

Drains all messages buffered for the current session since the last call.

Parameters:

• from (optional) — omit for Saved Messages only, "all" for all sources, or a chat ID/name to filter

telegram_wait

Efficient alternative to polling. Blocks until a message arrives or the timeout expires.

Parameters:

• timeoutMs (optional) — max wait in milliseconds; default 120000, max 300000

Use this when you want Claude to wait while you compose a reply:

"Wait for my next Telegram message before continuing."

telegram_login

Takes no parameters. Triggers a fresh Telegram authentication flow. Use this if the session has expired or credentials have been revoked.

telegram_contacts

Lists your Telegram contacts.

Parameters:

• search (optional) — filter by name or username
• limit (optional) — maximum results to return

telegram_chats

Lists your Telegram chats (groups, channels, and direct messages).

Parameters:

• search (optional) — filter by chat name
• limit (optional, default 50) — maximum results to return

Chat IDs returned here can be passed directly to telegram_history.

telegram_history

Fetches message history for a chat.

Parameters:

• chatId (required) — chat ID, username, or 'me' for Saved Messages
• count (optional, default 20) — number of messages to return (most recent first)

telegram_voice_config

Gets or sets the voice mode configuration. Configuration is persisted to ~/.telex/voice-config.json and survives watcher restarts.

Parameters:

• action (required) — 'get' to read current config, 'set' to update it
• defaultVoice (optional) — default voice name (e.g. 'bm_fable')
• voiceMode (optional, boolean) — true to enable automatic voice responses
• localMode (optional, boolean) — when true and voiceMode is true, use telegram_speak (Mac speakers) instead of telegram_tts (Telegram voice notes)
• personas (optional) — map of names to voice IDs (e.g. {"Nicole": "af_nicole", "George": "bm_george"})

Default personas: Nicole → af_nicole, George → bm_george, Daniel → bm_daniel, Fable → bm_fable

telegram_speak

Same TTS engine as telegram_tts, but plays audio through the Mac's speakers instead of sending a Telegram voice note. No Telegram connection required. Audio plays in the background without blocking other operations.

Parameters:

• text (required) — text to speak aloud
• voice (optional) — voice name (same list as telegram_tts); omit to use the configured default

telegram_rename

Renames the current Claude session (the iTerm2 tab name and session registry entry).

Parameters:

• name (required) — new session name

telegram_restart

Takes no parameters. Restarts the current Claude session: sends SIGTERM to the Claude process, waits for the shell prompt, then types claude to relaunch in the same directory.

telegram_discover

Takes no parameters. Scans all iTerm2 sessions for running Claude instances, updates the session registry, and prunes stale entries.

telegram_sessions

Takes no parameters. Returns the full list of registered Claude sessions with their names, indices, and active status.

telegram_switch

Switches incoming message routing to a different Claude session.

Parameters:

• target (required) — session index (1-based) or a substring of the session name

After switching, all incoming Telegram messages are delivered to the new session.

telegram_end_session

Closes a Claude session by sending SIGTERM and removing it from the registry.

Parameters:

• target (required) — session index (1-based) or a substring of the session name

telegram_command

Executes a Telegram slash command directly from Claude without a round-trip through the Telegram app. Useful for scripting session management.

Parameters:

• text (required) — command text, e.g. /sessions, /restart, /ss

Telegram commands

Messages sent from your phone starting with / are intercepted by the watcher:

| Command | Description | |---------|-------------| | /h or /help | Show help | | /s or /sessions | List open sessions (Claude and terminal) | | /N or /N name | Switch to session N (optionally rename) | | /n or /new | Create a new Claude session | | /c | Clear context and resume (/clear + go) | | /ss or /screenshot | Capture active session's iTerm2 window | | /cc | Send Ctrl+C (interrupt) | | /esc | Send Escape | | /enter | Send Enter/Return | | /tab | Send Tab (trigger completion) | | /up /down /left /right | Arrow keys | | /p | Pause session | | /pick N [text] | Select menu option N, optionally type text | | /restart | Restart the watcher |

Keyboard control

Send raw keystrokes to the active iTerm2 session from your phone:

| Command | Keystroke | Use case | |---------|-----------|----------| | /cc | Ctrl+C | Interrupt a running process | | /esc | Escape | Dismiss a dialog, exit a mode | | /enter | Return | Confirm a prompt | | /tab | Tab | Trigger shell completion | | /up | Up arrow | Previous history / menu up | | /down | Down arrow | Next history / menu down | | /pick N | Down x (N-1) + Enter | Select the Nth option in a menu |

Multiple sessions

Telex supports multiple simultaneous Claude Code windows. Each MCP server instance registers its TERM_SESSION_ID when it starts. The first session to register becomes the active recipient for incoming messages and stays active until explicitly switched or disconnected.

The watcher maintains a separate incoming message queue for each registered session. Use /N name from Telegram or telegram_switch from MCP to change which session receives incoming messages. Sending a message does not automatically change the active session.

Sessions register with a name derived from the working directory (e.g. a Claude session in ~/projects/myapp registers as myapp). Use /s to see all sessions and /N name to assign a custom name.

Routing is sticky: only the /N command changes the active session, not sending a message. This means switching sessions from your phone requires an explicit /N command.

Best practices for CLAUDE.md

To get the most out of Telex, add these rules to your CLAUDE.md (or ~/.claude/CLAUDE.md for global config):

Mirror every response to Telegram

Tell Claude to send the same content it prints on the terminal to Telegram, so you can follow along from your phone:

Every response you give on the terminal MUST also be sent to Telegram via telegram_send.
Send the same content — do not shorten or paraphrase.
Adapt markdown for Telegram: use **bold** and *italic* only. No headers, no code blocks.

Acknowledge before long tasks

If Claude is about to spawn agents, read multiple files, or do anything that takes more than a few seconds, it should send a brief Telegram message first — before calling any other tools. Otherwise your phone goes silent and you don't know if Claude heard you.

If a task will take more than a few seconds, your FIRST tool call must be
telegram_send with a brief acknowledgment (e.g. "On it — researching that now.").
Then proceed with the actual work. Never leave Telegram silent while working.

Drain the queue at session start

Messages you send from your phone while Claude is generating a response may be queued. Call telegram_receive early in each session to catch them:

At the start of every session, call telegram_receive to drain any queued
messages that arrived while you were offline.

Configuration

| Variable | Default | Description | |----------|---------|-------------| | TELEGRAM_API_ID | (required) | Telegram API ID from my.telegram.org | | TELEGRAM_API_HASH | (required) | Telegram API hash from my.telegram.org | | TELEX_TTS_VOICE | bm_fable | Default TTS voice |

CLI commands

# First-time setup: configure MCP and print authentication instructions
npx -y @tekmidian/telex setup

# Start the watcher daemon (manages iTerm2 delivery and IPC)
node dist/index.js watch

# Remove MCP config and stored credentials
node dist/index.js uninstall

Watcher daemon

The watcher is the core of Telex. It runs as a macOS launchd agent so it starts automatically and restarts if it crashes.

Starting manually

node dist/index.js watch

launchd setup (auto-start)

scripts/watcher-ctl.sh start    # Install and start as launchd agent
scripts/watcher-ctl.sh stop     # Stop and unload
scripts/watcher-ctl.sh status   # Show running state

The agent uses KeepAlive: true and ProcessType: Interactive. The Interactive process type and LimitLoadToSessionType: Aqua are required so the watcher can access the macOS GUI session and call AppleScript to control iTerm2.

Manual MCP configuration

Add to ~/.claude.json:

{
  "mcpServers": {
    "telex": {
      "command": "npx",
      "args": ["-y", "@tekmidian/telex"]
    }
  }
}

Using a local build:

{
  "mcpServers": {
    "telex": {
      "command": "node",
      "args": ["/path/to/Telex/dist/index.js"]
    }
  }
}

Troubleshooting

"Watcher not running" errors

The watcher daemon is not running. Start it with node dist/index.js watch or use scripts/watcher-ctl.sh start.

Telegram authentication failed

Ensure TELEGRAM_API_ID and TELEGRAM_API_HASH are set. Run node dist/index.js setup to re-authenticate.

Messages not appearing in Claude

Check that the watcher is running: ps aux | grep "telex.*watch". Verify the session ID matches your Claude tab.

"iTerm2 wants to control..." security prompt

Click OK. If you clicked "Don't Allow", go to System Settings > Privacy & Security > Automation and enable iTerm2.

TTS fails with "ffmpeg not found"

Install ffmpeg: brew install ffmpeg.

First TTS call takes a long time

The Kokoro model (~160 MB) is downloaded on first use and cached locally. Subsequent calls are fast.

Security

• Session credentials are stored locally in ~/.telex/. Treat them like passwords — they grant full access to your Telegram session.
• Telex reads and sends messages via your Saved Messages (self-chat) by default.
• No data is sent to any third-party service beyond Telegram's servers via MTProto.
• TTS synthesis is fully local (Kokoro-js runs on-device). Audio never leaves your machine.

Uninstall

node dist/index.js uninstall

Or with npx from the published package:

npx -y @tekmidian/telex uninstall

This removes Telex from ~/.claude.json and deletes credentials from ~/.telex/auth/. To fully clean up:

1. Stop the launchd agent: scripts/watcher-ctl.sh stop
2. Remove the plist: rm ~/Library/LaunchAgents/com.telex.watcher.plist (filename may vary — check scripts/watcher-ctl.sh for the exact name)
3. Run the uninstall command above to remove MCP config and auth
4. Optionally remove ~/.telex/ entirely: rm -r ~/.telex/

Restart Claude Code to apply.

Requirements

• Node.js >= 18
• aibroker — shared core (installed automatically)
• Telegram account with API credentials
• macOS with iTerm2 for the watch command and iTerm2 delivery
• ffmpeg for TTS voice note sending

License

MIT — see LICENSE

Author

Matthias Nott — github.com/mnott