AI & TechArtificial IntelligenceBigTech CompaniesNewswireTechnologyWhat's Buzzing

Meta: Muse resembles OpenClaw by design

▼ Summary

– Meta’s new personal agent, Muse, has faced accusations of being a clone of the open-source tool OpenClaw due to striking similarities in internal files.
– Nat Friedman, head of product at Meta Superintelligence Labs, confirmed that Muse was heavily inspired by OpenClaw and its creator, Peter Steinberger.
– Users identified near-identical content and file names, such as SOUL.md, leading Ansh Nanda to claim Muse is essentially OpenClaw repackaged for general users.
– Friedman explained that while built from scratch, the team intentionally adopted specific design choices like workspace naming conventions because they were optimal.
– OpenClaw remains popular among developers, with its framework appearing in other major tech products despite its creator joining OpenAI.

Meta’s head of product, Nat Friedman, has openly acknowledged that the company’s new personal agent, Muse, was heavily inspired by the open-source project OpenClaw. This admission comes after weeks of speculation and user comparisons suggesting that Meta’s latest tool bore a striking resemblance to its competitor. Friedman posted on X on Monday evening, stating, “We built Muse from scratch, but it is definitely heavily inspired as a product by OpenClaw.”

The controversy began when Ansh Nanda, co-founder of an AI app company, suggested that Muse was essentially OpenClaw repackaged for mainstream users. Nanda shared a conversation with Muse itself, where the agent confirmed that the similarity between its system files and those of OpenClaw was intentional. Users quickly identified specific parallels, including the presence of a SOUL.md file, which defines an agent’s personality, tone, values, and boundaries. The content within this configuration file was found to be nearly identical across both platforms.

Friedman clarified the team’s approach in his response. He revealed that he had used OpenClaw in January and subsequently purchased hundreds of Mac minis for his team at Meta Superintelligence Labs. Many team members became fans of the tool, prompting leadership to consider how they could build something similar that Meta could make safe, secure, and scalable for billions of users. When pressed on why Muse copied exact file names and near-identical content despite being built from scratch, Friedman did not dispute the premise. He explained, “The team thought Peter [Steinberger] had got those things exactly right.”

The Origins of OpenClaw

OpenClaw was created by Austrian developer Peter Steinberger and released as open-source software. It gained significant traction among developers in January after undergoing two previous iterations named Clawdbot and Moltbot. The project serves as a harness or scaffolding around a language model, managing instruction files, tool definitions, and the decision-making loop for agent actions. Steinberger joined OpenAI in February, and his framework continues to influence other tech giants, including Google Home, which recently opened its platform to third-party agents like OpenClaw.

Because OpenClaw is licensed under an open-source agreement, the file similarities raise product strategy questions rather than legal ones. Open source licenses are designed to allow others to read code and build upon it. Friedman, who previously served as CEO of GitHub before joining Meta alongside Chief AI Officer Alexandr Wang, is well-versed in these community norms. However, the licensing does not resolve the underlying product differences. Unlike OpenClaw, which has no central gatekeeper, Meta controls the connectors that integrate outside services into Muse, reviewing and approving which ones get featured.

Adoption and Market Impact

Meta has a history of adopting successful features from competitors, most notably Instagram Stories, which mirrored Snapchat’s 2013 launch. Mark Zuckerberg acknowledged this pattern during a US antitrust hearing in 2020. With Muse, Meta is applying similar tactics to the emerging field of personal AI agents. The app, launched on September 8, reached number one on the US App Store within days. Internal data reviewed by The Information indicates that over 500,000 people tried Muse in its first week, with more than 250,000 daily users submitting over two million prompts.

Muse is designed to handle tasks such as booking appointments, filling forms, managing customer service, making purchases, and generating content. It currently runs on macOS and integrates with WhatsApp, requiring users to grant access to email, calendar, messaging, and social accounts, as well as permissions for files, microphone, camera, and location. The service is free up to a limit, with premium tiers available at $20 and $100 per month for additional capacity. Users must be at least 18 years old to use the application.

Security Concerns and Commercial Integration

Despite its rapid popularity, Muse has faced immediate hurdles. Amazon blocked the agent from facilitating shopping on its site, citing failure to identify itself and violation of terms of use. Additionally, security researcher Patrick Wardle disclosed a zero-day vulnerability in the Mac app. Meta responded with a hotfix, asserting that the practical risk to users was low. These incidents contributed to a volatile market reaction, though Meta shares ultimately jumped 11% on Monday, marking its largest single-day gain since April 2025.

In response to inquiries from TechCrunch, a Meta spokesperson directed attention to Friedman’s post, noting that the company had nothing further to add. Meanwhile, Meta is expanding its ecosystem by opening its connector platform to external developers. Companies like Shopify and Stripe have already integrated their payment systems into the agent, creating a commercial layer that distinguishes Muse from the purely open-source nature of OpenClaw.

(Source: The Next Web)

Topics

ai product similarity 95% meta ai strategy 88% developer community reaction 85% open source licensing 82% tech industry dynamics 78%
Show More