The popular open-source AI assistant, originally known as Clawdbot, has undergone two rapid rebrands in January 2026, culminating in its current name, OpenClaw. This series of name changes comes amid intense viral adoption and growing scrutiny over security and trademark issues for the self-hosted AI agent.
Initially launched as Clawdbot by creator Peter Steinberger, the project gained rapid traction for its ability to perform real-world actions like managing calendars and sending messages via various chat applications. The assistant, which Steinberger built to explore human-AI collaboration, was named after the ASCII lobster associated with Claude Code.
On January 27, 2026, Clawdbot was forced to rebrand to Moltbot following a trademark dispute with Anthropic, the company behind the Claude AI models. According to Peter Steinberger, Anthropic cited trademark confusion with “Claude,” prompting the immediate change. This abrupt rebrand led to a chaotic period marked by account hijackings, crypto scams leveraging the old name, and exposed servers.
Just three days later, on January 30, 2026, the project announced its second rebrand to OpenClaw. This latest change was described by Steinberger as “deliberately calm,” with proper trademark searches, domain securing, and migration code implemented before the announcement. The new name, OpenClaw, explicitly signifies its open-source, community-driven, and self-hosted nature, while retaining a nod to its “lobster lineage.”
The shift to OpenClaw also signals a significant change in the project’s strategic posture, moving from what was once a “cool hack” to aspiring to be robust infrastructure. The latest release includes 34 security-related commits and machine-checkable security models, directly addressing the vulnerabilities that surfaced during its rapid growth. Previously, security researchers, including Bitdefender, had found exposed dashboards revealing API keys and conversation histories, and Cisco’s AI Threat and Security Research team identified critical vulnerabilities in OpenClaw’s “skills.”
OpenClaw is now explicitly model-agnostic, supporting not only Claude but also OpenAI, KIMI, and Xiaomi MiMo, reinforcing its positioning as a versatile AI assistant that users control. This move aims to reduce reliance on any single vendor and avoid implied endorsements or brand confusion.
The project’s rapid evolution highlights the challenges and complexities of open-source AI development, particularly concerning security and intellectual property in a fast-moving field.