Internal documents point to a late August or early September rollout, followed by Meta's Watermelon model in October.
By Ryan Merket · Published
Primary source: The Information
Why it matters #
Hatch would give Meta a direct AI subscription business while its planned WhatsApp platform could put third-party agents in front of a global consumer base.
Mark Zuckerberg's Meta is targeting late August or early September for the launch of Hatch, a consumer AI agent inspired by OpenClaw, and October for a new model internally called Watermelon, The Information reported on August 24th, citing internal documents.
The schedule puts a product and a business model behind Zuckerberg's push for "personal superintelligence." In an August 10th essay, Zuckerberg said Meta would build personal agents and models for billions of people, offer free versions and charge users who want more computing capacity. Hatch appears to be the first direct test of how much those users will pay.
A consumer wrapper around OpenClaw
OpenClaw is an open-source personal agent that can manage email, calendars and other services through messaging apps. Its software runs on a user's own machine and supports add-on capabilities called skills, but installing and configuring the system still requires command-line work and decisions about models, credentials and permissions.
Meta's pitch is easier packaging backed by its own infrastructure. According to the documents reviewed by The Information, Hatch has been trained to access services including DoorDash, Etsy, Reddit, Yelp and Outlook. Early prototypes include a customizable dashboard where users can see tools generated by the agent, such as a fitness tracker or travel itinerary.
The Information first reported the Hatch project in May, when Meta was still working toward internal testing. Hatch was then using Anthropic's Claude Opus 4.6 and Claude Sonnet 4.6 during development, with Meta planning to move the product onto its own Muse Spark models for launch, according to that report.
Meta is also preparing a WhatsApp platform that would let users connect other AI agents and converse with them through the messaging service. A limited rollout could begin during the week of August 24th, according to the latest report. That gives Meta a second route into the agent market: sell its own consumer agent while turning WhatsApp into a distribution channel for agents built elsewhere.
The $199.99 test
Final pricing has not been set. Meta has considered a tiered system with a premium Hatch subscription costing as much as $199.99 per month, The Information reported in June. Internal plans described a "Hatch Plus" tier with five to 10 times the daily capacity of the free product, with usage allowances resetting at the end of each billing cycle. A subscription at that level would move Meta beyond the mass-market pricing associated with consumer social apps. It would target professionals and heavy users willing to pay for an agent that can build small software tools and complete long-running tasks across multiple services.
The financial incentive is plain. Meta reported $60.8 billion in second-quarter revenue, including $59.36 billion from advertising. Ads therefore supplied roughly 98% of revenue during the quarter. Meta also expects 2026 capital expenditures, including finance lease payments, of $130 billion to $145 billion, nearly twice its 2025 spending range at the midpoint.
Hatch does not need to replace advertising to matter. A paid agent can give Meta recurring revenue tied directly to AI usage while helping offset the inference and infrastructure costs created by users who ask agents to browse websites, operate software and run multi-step workflows.
Watermelon follows a four-month model sprint
The October target for Watermelon extends a rapid release schedule from Meta Superintelligence Labs. Meta introduced Muse Spark on April 8th as a multimodal reasoning model with tool use and multi-agent orchestration. It followed with Muse Spark 1.1 on July 9th, adding stronger computer-use and coding capabilities.
On August 5th, Meta released Muse Spark 1.2 and the Muse Code terminal agent. Five days later, Meta published Muse Glimmer, a 30-billion-parameter open-weight model designed for local agent workflows.
The Information could not determine whether Watermelon will join the Muse family or ship under a separate name. The internal codename and October target also leave open which model will ultimately power Hatch as the agent reaches a broader audience.
Meta is pursuing open and proprietary releases in parallel. Glimmer gives developers weights they can run locally, while Hatch packages agent capabilities as a controlled consumer service with potential subscription limits. Zuckerberg's strategy depends on both: open models to keep developers building around Meta technology, and paid products that turn the resulting demand into revenue.
Meta already has the distribution through WhatsApp, Instagram, Facebook and its standalone AI app. Hatch will test whether it can convert that reach into a paid agent business before OpenAI, Anthropic and the open-source tools Meta is borrowing from establish the default way consumers delegate work to software.