Baidu reveals Xiaodu AI glasses “brain”, details 5 functions, cat's-eye models issued in December

Today, 2025BaiduWorld Congress opened in Beijing, at the conference, Baidu Group Vice President, Xiaodu Technology CEO Li Ying announced that theXiaodu voice assistant will be fully upgraded to multimodal AI intelligent assistant Super Xiaodu, all new Xiaodu products will be equipped with Super Xiaodu, and theTens of millions of Xiaodu sold devices support free upgrade to Super Xiaodu.

Super Xiaodu is a natural interaction multimodal AI intelligent assistant based on AI native operating system and covering the whole ecological scene.

Xiaodu AI Glasses Pro and Xiaodu Smart Camera, which are built on this capability, are both on sale.

Xiaodu AI glasses Pro is priced at 2,299 yuan and has been on sale on November 10, supporting call noise reduction, taking photos, shooting video, and AI literacy, AI translation, AI memo, ambient song list, and AI meeting minutes ability.

Xiaodu Smart Camera can record the screen and understand the scene and then intervene proactively based on specific behaviors preset by the user, such as pets in the kitchen, reminding the elderly to take their medication, and so on.

Ying Li talked about how Super Xiaodu can achieve from single to global, from passive to active understanding as well as personalized memory through multi-modal perceptual interactions such as text, sound, picture, vision and gesture. She believes that this is an important change in human-computer interaction, which can make the device change from a passive tool for executing commands to an active thinking partner.

First, Xiaodu AI glasses Pro is priced at 2,299 yuan, with glasses on hand to ask questions and records

Xiaodu AI Glasses ProIt has officially gone on sale on November 10th with a starting price of 2,299 RMB.

The glasses weigh 39 grams, come with a configurable nosepiece, a 12-megapixel wide-angle camera, a combined battery life of over 7.5 hours, a battery life of up to 68 hours with a rechargeable eyeglass case, and dual Bluetooth and Wi-Fi connectivity. Its lens solution offers photochromic lenses, which change color based on light and darkness.

The Xiaodu AI glasses are available in the original Boston frame model and the cat-eye model; the original Boston frame model is currently shipping and the cat-eye model goes on sale next month.

▲Boston original frame model (left), cat's eye model (right)

Specific features.Xiaodu AI glasses support call noise reduction, taking photos, and shooting videos. In terms of intelligent capabilities, Xiaodu AI glasses can realize AI literacy, AI translation, AI memo, ambient song list, AI meeting minutes.

AI recognitionHas become the basic function of AI glasses, the user's hand holding vegetables, the glasses can recognize the type of vegetables, and give the corresponding cooking advice.

existrenderingIn terms of this, Xiaodu AI glasses can start real-time broadcasting within 3 seconds, and Li Ying revealed that they have also optimized the thesaurus for dozens of fields in finance, medicine, and law, so that their translation accuracy and naturalness can be improved.

Users wearing the glasses can alsoreal time recordingParking, property information, etc., and subsequently be able to ask Xiaodu questions in real time to display parking information, set a reminder schedule, or directly call the phone number on the recorded memo message.

Xiaodu also linked up with NetEase Cloud Music to create theAmbient Song ListFunctions, such as the following demonstration, the user cycling at the beach, Xiaodu AI glasses automatically played the background music of "want to go to the beach".

endMinutes of the AI meetingFunction, when the meeting scene has a lot of intelligent hardware, glasses can be recorded naturally, and realize the meeting recording, transcription, summarization, after the meeting can help users to analyze the quality and efficiency of the meeting, analyze the attitude of the speakers as well as the potential points of contention.

If the customer does not present the final decision during the live communication, the AI glasses can also analyze their concerns based on the context of the meeting and give follow-up suggestions. In addition, Xiaodu AI Glasses can help users shoot PPT during the recording process and insert it at the right place in the meeting summary.

The AI minutes feature will go live in December.

Second, the smart camera supports AI care summarization, but also can link the whole house devices

Traditional caretaker cameras can record images, alarms, but its not understand what is happening behind the scenes, the maturity of multimodal large model technology allows the camera not only to recognize see the picture, understand the scene and even do active intervention.

Xiaodu smart camera is equipped with visual depth understanding of large modelsIts AI Attentive Care feature can alert users based on specific behaviors they set for people and pets, such as pets entering the kitchen, and will promptly alert users once the event is triggered again, or play the user's pre-recorded voice to actively intervene.

In addition, this camera supports whole-house device linkage, such as allowing speakers to play music in real time when a family member walks into the house.

Finally, Xiaodu smart camera can also generate AI care summary, for example, the user asked “today, the pet has entered the kitchen”, Xiaodu will generate voice broadcast summary, and automatically locate the corresponding location of the video.

Xiaodu released two smart cameras, Xiaodu Smart Camera C800 after the national complement to the arrival price of 365 yuan, C1200 for 399 yuan.

Third, “Xiaodu Xiaodu” is fully upgraded, has served 54 million families

“Xiaodu Xiaodu” voice assistant has been released so far has served more than 54 million families, along with the AI began to understand the real world, the positioning of the AI voice assistant has changed.

Super Xiaodu supports multimodal perceptual interaction, which enables understanding from single to global, from passive to active, and generates personalized memories for the user.

Li Ying cited an example, when the user asked Super Xiaodu “where is the remote control”, the camera will first look for the remote control in the home. The camera will first look for the remote control in the home, and if it doesn't find it, it will go back through 24 hours of surveillance footage to find where the remote control last appeared, and look at related videos to help the user recall the process of the remote control's disappearance.

Based on the super-powered Xiaodu, Xiaodu will also launch thecompanion intelligence, spending time with children, seniors and pets in the home.

Super Xiaodu can provide voice prompts when the child sits in an incorrect posture, adjust the lights when the study time is too long, remind the rest, but also summarize the child's study time, calculate its learning efficiency, etc., in the guise of a companion school caretaker.

In the face of the elderly, Super Xiaodu can remind them to take their medication, provide them with advice on dietary combinations, etc. If it monitors a pet's daily ear scratching, it will also allow owners to keep an eye on their pet's health.

Conclusion: Big Models are Accelerating Intelligent Hardware Upgrades and Brain Changes

Xiaodu announced this Xiaodu intelligent assistant upgrade, from smart speakers, smart cameras to smart glasses, to complete the whole line of products “brain”. Previously, the hardware devices represented by smart speakers are mostly in the primary stage of the Internet of Everything, and most of the devices are simply connected to each other and the instructions are executed, and now it is equipped with a multimodal model that can completely reshape the interaction experience between the user and the intelligent assistant.

From the layout of Xiaodu, it can be seen that it is promoting the intelligent upgrading of equipment, but also through the super Xiaodu to enhance the intelligent assistant's ability to understand complex needs and fuzzy instructions, and can link the whole house equipment to enhance its autonomy.

This may also mean that, along with the development of big model technology, the intelligent hardware track is ushering in new variables, and the intelligent upgrading of equipment and the synergistic linkage of multiple devices may become the key.

Article source: Wisdom