Code from anywhere. Control your coding CLI from any messaging app.

Send a WhatsApp message. Watch your coding agent write code.

txtcode bridges messaging apps to AI coding assistants, giving you full remote control over your development environment from your phone, tablet, or any device with a chat app.

Why txtcode?

Code without a keyboard. Send a message from your phone, and txtcode dispatches it to an AI coding agent running back at your desk. Describe a bug fix in WhatsApp, telegram, discord or signal and watch Claude Code patch it or Ask for tests, get them written by Codex. No laptop required.:

1. Text your AI from WhatsApp, Telegram, Discord, Slack, Teams, or Signal
2. It writes code using Claude Code, Cursor, Codex, Gemini CLI, or other adapters
3. You stay in control with mode switching and session logs

No port forwarding. No VPN. Just message and code.

✨ Features

Messaging-First

Connect via 6 platforms : WhatsApp, Telegram, Discord, Slack, Microsoft Teams, and Signal.

10 AI Providers

Anthropic, OpenAI, Google Gemini, Mistral, Moonshot, MiniMax, xAI Grok, Z.ai (Zhipu), HuggingFace, and OpenRouter. Instantly switch providers with /switch directly from your messaging app.

Hot-Switching

Use /switch to change AI providers or coding adapters on the fly without restarting the agent perfect when you hit rate limits or want to try a different model.

7 Coding Adapters

Claude Code, Cursor CLI, OpenAI Codex, Gemini CLI, Kiro CLI, OpenCode, and Ollama (local/free). Full coding CLI control in /code mode.

Session Logging

Per-session logs accessible from the TUI. Follow live, view by index, auto-pruned after 7 days.

🚀 Quick Start

Requires Node.js 20+ and npm 10+.

# Install globally
npm install -g txtcode-ai

# Launch the interactive TUI
txtcode

Or run without installing:

npx txtcode-ai

That's it. The interactive menu guides you through everything authentication, configuration, starting the agent, viewing logs, and more. No subcommands to memorize.

🐳 Docker

Run txtcode in a container without installing Node.js or npm locally.

Build the image:

docker build -t txtcode .

Run the container:

docker run -it \
  -v $(pwd):/workspace \
  -v ~/.txtcode:/root/.txtcode \
  txtcode

| Flag | Purpose | | :----------------------------- | :-------------------------------------------------- | | -v $(pwd):/workspace | Mounts your project directory into the container | | -v ~/.txtcode:/root/.txtcode | Persists config, session data, and logs across runs |

Note: API keys are stored securely via your OS keychain when running natively. Inside Docker, txtcode uses an encrypted file-based fallback (TXTCODE_DOCKER=1 is set automatically). You can also pass keys as environment variables with -e, e.g. -e ANTHROPIC_API_KEY=sk-....

📝 Supported Platforms

| Platform | Transport | Setup | | :----------------------------------------------------------------------------------------------------------- | :-------------- | :-------------------------------------------------------------------------------------- | | | QR code pairing | Scan QR in terminal on first run | | | Bot API | Create bot via @BotFather, paste token | | | Bot gateway | Create app at discord.com/developers, paste bot token | | | Socket Mode | Create app at api.slack.com, enable Socket Mode | | | Bot Framework | Register bot at dev.teams.microsoft.com | | | signal-cli REST | Run signal-cli-rest-api via Docker |

🤖 AI Providers

txtcode supports 10 LLM providers for chat mode. Configure one or more during setup and hot-switch with /switch.

| Provider | Example Models | Notes | | :------------------ | :----------------------------------------- | :-------------------------- | | Anthropic | claude-sonnet-4-6, claude-opus-4-6 | Claude family | | OpenAI | gpt-5.2, o4-mini, gpt-4o | GPT and o-series | | Google Gemini | gemini-2.5-pro, gemini-2.5-flash | Gemini family | | Mistral | mistral-large-latest, codestral-latest | Mistral + Codestral | | Moonshot (Kimi) | kimi-k2.5, moonshot-v1-128k | Long-context models | | MiniMax | MiniMax-M2.5, MiniMax-M2.1 | MiniMax family | | xAI (Grok) | grok-4, grok-3-fast | Grok family | | Z.ai (Zhipu) | glm-5, glm-4.7, glm-4.6 | GLM family | | HuggingFace | Discovered at runtime | Inference Providers API | | OpenRouter | Discovered at runtime | Unified API for 100+ models |

All providers are used in chat mode for general conversation and coding questions.

💻 Coding Adapters

Use /code mode to route messages directly to a coding adapter with full coding CLI control.

| Adapter | Backend | CLI Required | Notes | | :--------------------- | :------------- | :----------- | :-------------------------- | | Claude Code | Anthropic API | claude | Official Claude CLI | | Cursor CLI | Cursor | cursor | Headless Cursor | | OpenAI Codex | OpenAI API | codex | OpenAI's coding agent | | Gemini CLI | Google AI API | gemini | Google's CLI | | Kiro CLI | AWS | kiro-cli | AWS Kiro subscription | | OpenCode | Multi-provider | opencode | Open-source, multi-provider | | Ollama Claude Code | Local (Ollama) | ollama | Free, no API key needed |

💬 Chat Commands

Send these commands in any messaging app while connected:

| Command | Description | | :----------- | :---------------------------------------------------------------------------- | | /chat | Switch to Chat mode to send messages to primary LLM (default) | | /code | Switch to Code mode to send messages to coding adapter (full CLI control) | | /switch | Switch primary LLM provider or coding adapter on the fly | | /cli-model | Change the model used by the current coding adapter | | /cancel | Cancel the currently running command | | /status | Show adapter connection and current configuration | | /help | Show available commands |

⚙️ Configuration

Config is stored at ~/.txtcode/config.json. API keys and tokens are stored in your OS keychain (via keytar), never in the config file.

To modify settings, select Configuration from the main menu. Options include:

• Change Messaging Platform
• Change Coding CLI Type
• Change AI Provider
• Change Project Path
• View Current Config

🏷️ Logs

• Location: ~/.txtcode/logs/
• Format: session-YYYY-MM-DD-HHmmss.log (one per agent session)
• Cleanup: Files older than 7 days are pruned automatically

Verbose and debug output goes to the log file; the terminal shows only key status lines.

Troubleshooting

• Restart your terminal after npm install -g txtcode
• Ensure the global bin directory is in your PATH: npm config get prefix

• Use a stable internet connection
• Select Logout from the main menu, then Start Agent again to get a fresh QR
• Session is stored under ~/.txtcode; don't delete it to keep the same linked device

• Confirm bot token via Configuration → View Current Config
• Telegram: ensure the bot is not blocked and you've started a chat with it
• Discord: ensure MESSAGE CONTENT INTENT is enabled in the developer portal
• Slack: ensure Socket Mode is enabled and event subscriptions are configured

• Check API keys and credits for your provider
• Code mode: ensure the CLI for your adapter is installed and in PATH (claude, codex, gemini, kiro-cli, ollama, cursor, opencode)
• Check adapter and connection state via the main menu
• Select View Logs or check ~/.txtcode/logs/ for detailed errors