A lightweight, multi-channel AI assistant designed for 24/7 personal assistance, TinyClaw integrates with Discord, WhatsApp, and Telegram, supporting both Anthropic Claude and OpenAI GPT models. Its primary benefit is robust, sequential message processing via a file-based queue system, which prevents race conditions and maintains a unified conversation context across all connected platforms.

How It Works

TinyClaw utilizes a file-based queue (incoming/ and outgoing/) to manage message flow between its various channel clients (Discord, WhatsApp, Telegram) and the AI provider. Each incoming message is written to a file in the incoming/ queue. A dedicated processor then picks up messages one at a time, routes them to the configured AI (Anthropic Claude or OpenAI via CLI), and writes the response to the outgoing/ queue. This sequential, file-driven approach guarantees that messages are handled without race conditions, ensuring a consistent conversational state across all channels.

Quick Start & Requirements

• Prerequisites: macOS or Linux, Node.js v14+, Bash 4.0+ (macOS users may need to install/use a newer version via Homebrew), tmux, Anthropic Claude Code installed (for Anthropic provider), or Codex CLI installed and authenticated (for OpenAI provider).
• Installation: Clone the repository, navigate to the directory, and run npm install.
• Running: Execute ./tinyclaw.sh start. The first run initiates an interactive setup wizard to configure channels, AI provider, models, and API tokens.
• Setup: Requires Discord and Telegram bot tokens, and authentication for the Codex CLI. WhatsApp integration involves scanning a QR code.
• Links: Discord Developer Portal, Telegram BotFather.

Highlighted Details

• No Race Conditions: Employs a file-based queue system for sequential message processing, preventing concurrent access issues.
• Multi-Channel Support: Seamlessly integrates Discord, WhatsApp, and Telegram, all sharing the same conversation history.
• Flexible AI Providers: Supports Anthropic Claude (Sonnet, Opus) and OpenAI GPT models (e.g., GPT-5.3 Codex, GPT-4o), with easy switching via CLI commands.
• Persistent Sessions: Maintains WhatsApp session state locally for auto-reconnection; other channels rely on bot tokens.

Maintenance & Community

The provided README does not detail specific maintainers, community channels (e.g., Discord, Slack), sponsorships, or a public roadmap. It notes inspiration from OpenClaw.

Licensing & Compatibility

Licensed under the MIT License. This permissive license is generally compatible with commercial use and linking within closed-source projects.

Limitations & Caveats

Setup requires specific command-line tools (claude -c -p, codex exec) to be installed and properly authenticated, which may present a barrier. macOS users must ensure they are using a Bash version of 4.0 or higher, as the default version is insufficient. WhatsApp integration necessitates an initial QR code scan and the availability of a phone. Information regarding active project maintenance or community support is not explicitly provided in the README.