Meta launches its Meta AI app with memory capabilities, personalization, image generation, and voice mode

Meta has just presented the first version of the Meta AI app, its proprietary artificial intelligence application powered by its multimodal model Llama 4. The company assures that its users will be able to enjoy a highly personalized experience where the voice mode will play a significant role. Likewise, the social component that characterizes Meta’s other services will also be present in this app.

Meta AI app is now available for download in the App Store and on Google Play Store. However, some of its features have only been activated in certain countries such as the United States.

The Meta AI app is coming to 36 countries globally! All of them will have text, but the US, Australia, Canada and New Zealand will get first access to both text & voice capabilities!

— Meta (@Meta) April 29, 2025

What is the Meta AI app like?

This new app with which Meta seeks to gain a foothold against ChatGPT is the evolution of its digital assistant Meta AI, an AI launched in September 2023 and which, until now, was only accessible in certain regions via Meta platforms or as a chatbot on its own website.

• To learn more: Meta is developing the Meta AI app, an AI application with which to compete against ChatGPT

Meta AI app allows for enriched conversations with this AI assistant, thanks to its ability to remember your preferences and context. In addition, its chatbot design and search capabilities on the web allow for easy and fluid use of the tool.

It should be noted that this novelty has been launched leveraging the structure of Meta View, the app that allowed managing Meta’s AI glasses. Users of this application, now revamped into Meta AI app, will be able to continue managing their devices from the Devices tab.

Voice mode

Thanks to the enhanced technology of Llama 4, Meta AI app will be able to provide more personal and useful responses in its voice mode while employing a more fluid and conversational style. Additionally, it will also be possible to request the AI to generate or edit images through spoken conversations.

On another note, Meta has included a duplex voice demonstration to conduct conversations in which you can speak and the AI also responds using voice instead of merely reading written responses. Although, in this case, the AI does not have real-time internet access, something that is expected to be resolved later. This function can be activated or deactivated as desired.

For now, the voice mode and its experimental duplex feature are only available in the United States, Canada, Australia, and New Zealand.

Intelligence and Memory

You can ask the Meta assistant to remember certain key data about you, your likes and interests, and even instruct it to gather important information according to the context. Additionally, Meta AI will also rely on the data you have shared in other Meta products or content you have interacted with to ensure more personalized responses and results.

These capabilities have already been activated for users in the United States and Canada.

Feed Discover

To continue promoting connections among people, the tech company has decided to integrate a Feed called Discover in Meta AI app. This is a space where users can explore how others are utilizing this AI assistant and share their own experience if they so wish.

Photo: Meta