What is Hatch and why it matters

According to a report by The Indian Express from June 8, 2026, Meta is intensively testing a new AI agent codenamed Hatch. This isn't another ChatGPT-style chatbot that just answers questions. Hatch is meant to be an autonomous agent — a system that independently performs tasks based on natural language instructions.

In the test version, Hatch can already create custom software tools based on a verbal description. A user says "build me a fitness tracker" and Hatch writes the corresponding application. According to available information, the agent works similarly to OpenClaw, a tool that gained attention this year for its ability to automate digital tasks locally on the user's device.

What Hatch is supposed to do

Leaked information suggests that Hatch will have several key features:

• Tool management dashboard — an overview page where the user can see all apps and automations that Hatch has created
• Modular skills — expandable capabilities for workflow automation, integration with third-party services, and specialized tasks like planning travel itineraries
• Background task execution — sending emails, scheduling meetings, creating presentations with minimal manual input
• Integration with the Meta ecosystem — primarily with Instagram, giving Hatch potential reach to billions of existing users without the need for separate registration

A price that levels the playing field

The most surprising aspect of the entire report is the pricing strategy. Meta is reportedly considering a two-tier subscription:

• Hatch (free version) — limited number of daily requests
• Hatch Plus — approximately $200 per month (roughly 4,600 CZK), with significantly higher limits

This places Hatch alongside ChatGPT Pro from OpenAI and Claude Max from Anthropic — both also at $200 per month. It's a clear signal that Meta doesn't want to compete on price, but on capabilities. For comparison: ChatGPT Plus costs $20/month, Claude Pro also $20/month, while the premium versions with higher limits and agent features cost ten times as much.

For Czech users and businesses, this means that premium AI agents are becoming the standard — not a luxury for enthusiasts. The question remains whether Hatch Plus will be available in the Czech Republic immediately at launch, or whether Meta will deploy gradual geographical expansion as with previous products.

Technical background: From Claude to Muse Spark

An interesting technical detail is that Meta is currently testing Hatch with Anthropic's Claude Sonnet models — that is, with technology from a direct competitor. The final product, however, should run on Muse Spark, Meta's own AI model, which the company has not yet presented in detail. This approach is not unusual in the industry: companies often test concepts on third-party models and only then migrate to their own infrastructure.

Meta has lagged behind the leaders in language models. Its Llama 4 models (open-source) are among the most popular alternatives to GPT and Claude, but they aren't at the top in raw benchmark results. Betting on agents — autonomous systems that don't just answer but act — represents a strategic pivot: Meta doesn't want to win the race in parameter count, but in practical usefulness.

Comparison with competitors

Several players already operate in the AI agent market today:

• OpenAI — ChatGPT with features like Code Interpreter, automatic web browsing, and task planning. The premium ChatGPT Pro plan ($200/month) offers unlimited access to all models including GPT-5.
• Anthropic — Claude Max ($200/month) with expanded limits and the agent feature "computer use," where Claude can control a computer like a human — move the cursor, click, type.
• Google — Gemini Advanced with Google Workspace integration ($20/month), but without a standalone $200 agent plan.
• xAI (Elon Musk) — Grok with a premium plan for X (Twitter), but without full-fledged agent functionality yet.

Hatch differentiates itself particularly through integration into social platforms. While ChatGPT or Claude require a separate app or web interface, Hatch could be present where people already spend their time — in Instagram, WhatsApp, or Messenger. This distribution advantage is crucial: Meta has over 3 billion active users across its platforms.

What Hatch means for regular users

For a Czech user who today uses AI mainly for writing emails, translations, or summarizing texts, the rise of agents represents a fundamental shift. It's no longer about AI "answering." It's about AI "doing."

Imagine giving Hatch this instruction: "Plan me a weekend trip to Vienna, find a train connection from Brno, book accommodation under 3,000 CZK per night, and prepare an itinerary based on my interest in modern architecture." If Hatch meets expectations, it eliminates an hour of clicking between search engines, maps, and booking portals.

At the same time, caution is warranted. An agent with access to your accounts, calendar, and payment methods presents a security challenge. Meta has not yet disclosed how it will address privacy and data security issues — and given its controversial history with user data processing (see the Cambridge Analytica scandal), trustworthiness will be a key factor.

The European dimension: EU AI Act and availability in the Czech Republic

For European users, the question of regulation is essential. The EU AI Act, effective from 2025, classifies AI systems by risk. AI agents that autonomously perform actions on behalf of the user (such as financial transactions or access to personal data) fall into the high-risk category. This means stricter requirements for transparency, testing, and documentation.

Meta will need to ensure for the European market:

• Transparent labeling of AI-generated content and actions
• GDPR compliance in processing personal data
• Language choice — ideally including Czech

Whether Hatch will support Czech at launch is not yet known. Meta AI (the chatbot in Messenger and WhatsApp) handles Czech reasonably well, so the chances of localization are real. For Czech businesses, Hatch could represent an interesting alternative to ChatGPT — especially if it's integrated into tools that employees already use.

What to expect next

A launch date for Hatch has not yet been set. Meta is still testing the product and the final version may differ from current leaks. What is certain is that Zuckerberg sees AI as the main engine of future revenue. Just in recent months, Meta expanded its premium Meta AI subscription, reorganized internal teams around AI development, and is massively investing in computing infrastructure.

For readers of jarvis-ai.cz, we will be following the developments. If Hatch appears in the Czech environment, we will bring a detailed review and comparison with competitors.