Meituan Releases AI Browser Tabbit 1.0
Meituan’s GN06 team has launched Tabbit 1.0, an AI-native browser that can carry out complex tasks across apps and web pages. The new version adds a memory feature that stores user preferences and context.
Intelligence analysis by GPT-5.4 Mini
Tabbit is positioned as an AI entry point in browser form, with multiple large models built in. Users can type a request and the browser is supposed to handle cross-software and cross-web tasks, while the 1.0 release adds persistent memory for preferences, background, and other important details.
Meituan made a browser that acts more like a helper than a normal window for websites. It can remember what someone likes and try to do bigger jobs across different apps and pages, like a smart assistant that keeps notes and follows them.
Analysis
What Meituan announced
On June 9, Meituan’s GN06 team said Tabbit 1.0, an AI-native browser, has officially launched. The product is described as a browser-shaped AI entry point rather than a conventional browsing tool.
According to the announcement, Tabbit has multiple large models built in. A user can enter a request, and the browser is intended to automatically carry out complex tasks across software and web pages. That framing puts it in the category of agent-like tools that try to do work for the user instead of simply answering questions.
What is new in 1.0
The main new feature highlighted in the release is memory. Tabbit 1.0 will continuously record user preferences, background, and other important information, then turn that into “callable memory.” The product is also meant to adapt replies to the user’s style and reduce unnecessary conversation and actions.
How the article frames the product
The piece is a short announcement rather than a detailed product review. It does not include pricing, availability details beyond the launch claim, or technical benchmarks. The key point is that Meituan is presenting Tabbit as an AI front door for tasks, with memory as the differentiating feature.
For China’s AI and product scene, the story fits a broader move toward agentic interfaces: software that remembers context, accepts natural-language instructions, and tries to complete multi-step work across different apps and sites instead of forcing users to click through each step manually.
Key points
- Meituan’s GN06 team says Tabbit 1.0 is now live.
- Tabbit is described as an AI-native browser and an AI entry point.
- The browser includes multiple large models and is meant to execute complex cross-app and cross-web tasks.
- Version 1.0 adds a memory feature that records preferences, background, and other important information.
- The product aims to reduce unnecessary dialogue and actions by adapting to the user’s style.
If Tabbit works well, it could make complicated online tasks faster and easier by reducing repeated typing and clicking. The memory feature could also make the browser feel more personal and useful over time, since it can remember preferences and context.
The article does not show whether Tabbit can reliably complete complex tasks in real use, so the feature set may be more ambitious than practical. A browser that stores personal context also raises user-trust and privacy concerns, especially if the memory is not transparent or easy to control.
