Meta to launch paid AI agent Hatch, Watermelon model in Oct

Meta Platforms announced that it will introduce a consumer‑focused AI assistant called Hatch within the next few weeks, and that a new underlying model named Watermelon is scheduled for release in October. The information comes from internal documents reviewed by The Information and reported by The Decoder. Hatch is positioned as a paid service that will allow users to interact with a range of third‑party sites, extending Meta's AI capabilities beyond its existing ecosystem.

The rollout plan includes a tiered pricing structure, with a premium subscription option that could cost as much as $199.99 per month. Hatch is designed to pull data from services such as DoorDash, Etsy, Reddit, Yelp, and Outlook, providing a customizable dashboard where users can add widgets like a fitness tracker or a trip planner. By bundling these integrations into a single interface, Meta hopes to create a more compelling value proposition for consumers willing to pay for AI‑driven convenience.

This product launch is part of CEO Mark Zuckerberg's broader strategy to monetize the company's substantial investment in artificial intelligence and to diversify revenue away from its traditional advertising base. The internal documents suggest that Hatch will serve as a consumer version of Meta's internal OpenClaw agent, indicating a shift from internal tooling toward a marketable offering for everyday users.

OpenClaw has previously been used within Meta for internal tasks, but the company is now repackaging that technology for a broader audience. By offering a paid subscription, Meta joins other AI firms that have introduced premium tiers for their models, aiming to capture a slice of the growing market for AI‑enhanced personal assistants. The pricing ceiling of nearly $200 a month signals that Meta is targeting power users and businesses that can justify the expense for advanced features and integrations.

The upcoming Watermelon model, slated for an October launch, adds another layer of uncertainty. It is not yet clear whether Watermelon will be incorporated into the existing Muse family of models or released as a standalone architecture. This ambiguity leaves developers and potential customers waiting for more technical details about performance, size, and the specific capabilities that will differentiate Watermelon from Meta's current offerings.

In parallel with Hatch, Meta is also developing a platform on its WhatsApp messaging service that will let users plug in additional AI agents. This move could turn WhatsApp into a hub for AI‑driven interactions, allowing third‑party developers to create specialized agents that operate within the familiar chat environment.

If Hatch successfully integrates the listed services and delivers a smooth user experience, it could become a central point for daily digital tasks, potentially reducing the need for users to switch between multiple apps. However, the high subscription price and the lack of publicly available benchmark data may limit early adoption, especially among cost‑conscious consumers.

Stakeholders should monitor the official launch dates, the final pricing tiers, and any performance metrics that Meta releases for the Watermelon model. The degree to which the WhatsApp platform opens up to third‑party agents will also be a key indicator of how broadly Meta intends to expand its AI ecosystem. Future updates are expected to clarify whether Watermelon will belong to the Muse family or stand alone, and how it will compare to competing models in terms of speed, accuracy, and resource requirements.