.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
In less than five weeks, one of the fastest-growing open-source AI projects changed its name twice. Clawdbot became Moltbot. Moltbot became OpenClaw. The code barely moved, but the branding story matters because it explains legal pressure, open-source risk and why the project looks more stable now than it did at launch.
Clawdbot: Explosive growth meets trademark reality
Clawdbot started as a personal project by Peter Steinberger, an Austrian developer best known for founding PSPDFKit. What began as a local experiment quickly turned into a local-first AI agent that could run on your own hardware and connect to platforms like WhatsApp, Telegram, and Discord.
Growth was immediate and intense. Within days of being open-sourced, the project surged on GitHub and drew massive attention across developer communities. Business press later reported millions of site visits in a single week and star counts rising at a pace rarely seen for infrastructure tools.
The name, however, became a problem. Clawdbot’s mascot and naming leaned heavily on the “Claude with hands” idea. On January 27, 2026, Anthropic contacted Steinberger over trademark concerns related to the visual and phonetic similarity between “Clawd” and “Claude.” Although the project was a legitimate API consumer, the trademark itself was enforceable.
Steinberger complied quickly and publicly. Clawdbot was retired as a name.
Moltbot: A temporary name with unintended fallout
The replacement name, Moltbot, was announced the same day. The logic was biological rather than legal. Lobsters molt to grow, and the project kept the lobster theme while shedding the problematic branding.
The name was chosen fast, during a chaotic community discussion, and it never fully settled. Two days later, a second issue surfaced that had nothing to do with trademarks and everything to do with speed.
During the rename, there was a brief gap while old social and repository names were released and new ones were being claimed. In that window, scammers moved quickly. A fake token using the abandoned Clawdbot name appeared on Solana and briefly reached a reported eight-figure market cap before collapsing. Impersonation accounts spread across social platforms, and users were warned to avoid unofficial repositories and extensions.
Coverage from outlets like Business Insider and Yahoo Finance documented the confusion and the speed at which bad actors took advantage of the transition.
Internally, Steinberger later acknowledged that Moltbot never felt like a permanent identity. The community largely agreed.
OpenClaw: a deliberate reset
On January 29, 2026, the project was renamed again. This time, the change was voluntary.
OpenClaw combined two ideas the maintainers wanted to emphasize going forward: open-source development and continuity with the original claw motif, without leaning on any single AI vendor’s branding. Trademark research was done in advance, social handles were secured before the announcement, and the rollout was intentionally calm.
The official site moved to openclaw.ai, and the project repositioned itself as a model-agnostic agent framework rather than a Claude-adjacent tool.
What did not change technically
Despite the naming chaos, the underlying system stayed the same. All three names refer to the same codebase and architecture.
The core remains a local-first AI agent built on Node.js 22+, using a gateway and daemon model with a skills plugin system. The REST API still binds on port 18789. Configuration structure and flags are unchanged.
Existing installations migrated automatically. Configuration directories moved from ~/.moltbot/ to ~/.openclaw/ without user intervention, and legacy commands continue to function through compatibility shims.
From an operator’s perspective, this was a branding rebrand, not a fork.
What actually changed
The OpenClaw release bundled real improvements alongside the rename.
New model support was added, including KIMI K2.5 and Xiaomi MiMo-V2-Flash. New integrations appeared for platforms like Twitch and Google Chat. Image handling was introduced in the web chat.
Security also tightened. After scans revealed many publicly exposed instances bound incorrectly to all interfaces, the project added audit tooling and guardrails. A new openclaw security --audit command helps identify unsafe configurations, and public bindings now trigger stricter checks rather than silently exposing services.
The official onboarding and update path was consolidated around the installer hosted on the OpenClaw website, which detects existing installs and upgrades them in place.
Why this rebrand matters
This sequence is a case study in how fast open-source can move, and how fragile branding can be when growth outpaces legal and operational preparation.
Trademark enforcement is now common in the AI space, especially where model names are involved. Rapid renames create supply-chain risk, as seen in the scam token incident. At the same time, the episode shows that a strong technical foundation can survive branding turbulence.
OpenClaw’s current positioning avoids vendor lock-in, emphasizes self-hosted control, and reflects lessons learned under pressure. The name may be new, but the project beneath it is no longer experimental in the same way it was at launch.
For teams running AI agents on their own infrastructure, that distinction matters more than the logo ever did.
