The open-source personal AI assistant formerly known as Clawdbot has gone viral after rebranding to Moltbot, drawing attention for its always-on agentic behavior and deep system access. According to a TechCrunch report, Moltbot operates locally on users’ machines and can autonomously take actions, monitor tasks, and communicate updates through platforms like Telegram and WhatsApp.

The project was renamed after Anthropic raised trademark concerns over the original name Clawdbot. Creator Peter Steinberger rebranded the assistant as Moltbot, framing the change as symbolic of growth — “molting” into a new shell while keeping the same core capabilities.

Viral demonstrations show Moltbot negotiating purchases, booking services, and even making phone calls via integrations such as ElevenLabs. Unlike typical chat-based assistants, Moltbot is designed to run continuously in the background, maintain long-term context, and proactively execute tasks without repeated user prompts.

However, security experts have warned that Moltbot’s full system access introduces significant risks. Because it can read messages, access credentials, and control applications, a single exploit or prompt injection could potentially compromise an entire device if safeguards are not carefully configured.

Why it matters: Moltbot represents a major leap toward truly autonomous personal AI agents that “actually do things,” not just respond to prompts. At the same time, it highlights the growing tension between utility and security, as agentic systems blur the line between helpful automation and systemic risk.