ClawBlink: Create background AI agents from Telegram or WhatsApp chat

I built ClawBlink as a small experiment around a simple idea: could chat be the interface for automation?

Instead of dashboards or workflow builders, the user just sends a message in Telegram or WhatsApp describing what they want.

For example:

"Send me an AI news digest every morning"

That message gets converted into a small agent configuration which runs in the background.

Agents are simple pipelines such as:

http_request → llm_analyze → notify

The project is intentionally small (~1.9k lines of Python) and designed to be easy to understand and extend.

It currently supports Telegram bots and a WhatsApp Web bridge, and can use Ollama, Gemini, or OpenAI-compatible APIs.

I'm curious whether people prefer this kind of chat-driven automation compared to traditional dashboard-based tools.