An open supply AI assistant known as Moltbot (previously “Clawdbot”) lately crossed 69,000 stars on GitHub after a month, making it one of many fastest-growing AI initiatives of 2026. Created by Austrian developer Peter Steinberger, the device lets customers run a private AI assistant and management it by messaging apps they already use. Whereas some say it feels just like the AI assistant of the long run, working the device as at the moment designed comes with severe safety dangers.
Among the many dozens of unofficial AI bot apps that by no means rise above the fray, Moltbot is maybe most notable for its proactive communication with the person. The assistant works with WhatsApp, Telegram, Slack, Discord, Google Chat, Sign, iMessage, Microsoft Groups, and different platforms. It might probably attain out to customers with reminders, alerts, or morning briefings primarily based on calendar occasions or different triggers. The venture has drawn comparisons to Jarvis, the AI assistant from the Iron Man movies, for its capacity to actively try and handle duties throughout a person’s digital life.
Nevertheless, we’ll inform you up entrance that there are many drawbacks to the still-hobbyist software program: Whereas the organizing assistant code runs on a neighborhood machine, the device successfully requires a subscription to Anthropic or OpenAI for mannequin entry (or utilizing an API key). Customers can run native AI fashions with the bot, however they’re at the moment much less efficient at finishing up duties than the perfect industrial fashions. Claude Opus 4.5, which is Anthropic’s flagship massive language mannequin (LLM), is a well-liked selection.
A screenshot of speaking with Clawdbot / Moltbot taken from its GitHub web page.
Credit score:
Moltbot
Establishing Moltbot requires configuring a server, managing authentication, and understanding sandboxing for even a slice of safety in a system that mainly calls for entry to each aspect of your digital life. Heavy use can rack up vital API prices, since agentic programs make many calls behind the scenes and deplete a variety of tokens.
