Google released an update with two feature additions on July 16: AI Mode can now connect to third-party apps such as Instacart, Canva, and YouTube, converting AI search answers directly into in-app actions; Google Vids also adds a digital virtual avatar feature based on users’ selfies and voice recordings, with the avatar image featuring an invisible SynthID watermark and linked to the user’s Google account.
Google AI Mode App Connections
（Source: Google blog）
According to an official Google statement, in AI Mode, after linking to a related account, AI answers from search conversations can be directly converted into real actions within third-party apps. Google shared the following three demonstration scenarios:
Instacart: Users create a barbecue shopping list in AI Mode; ingredients are added directly to the Instacart cart, and checkout is completed
YouTube Music: AI Mode creates a party playlist and saves it to the user’s YouTube Music library
Canva: AI Mode creates a flyer template for the user
This feature is currently only available in the United States. Google says it is expanding its partner ecosystem and will soon support more applications. This AI Mode update builds on capabilities demonstrated at Google I/O, where Gemini app linking for Canva, OpenTable, Spark, and Instacart was already shown.
Google Vids Selfie Avatar Feature: SynthID Watermark Integration
Google is updating its Vids tool, allowing users to upload selfies and voice recordings to create a digital virtual avatar whose appearance and voice resemble the user. According to Google’s announcement, each virtual avatar includes an invisible SynthID watermark and is associated with the account holder’s Google account; in certain regions, access is limited to users aged 18 and above.
Google is also integrating Gemini Omni into Vids. Omni is Google’s multimodal model that supports combining text prompts with reference images to generate videos, and offers editing functions such as background replacement, brightness adjustments, and adding effects. This update also adds step-by-step editing support, allowing users to adjust the video without starting over. Vids was originally positioned as an AI-assisted work presentation tool; after this update, it expands into a general-purpose video creation platform.
ChatGPT External App Integration Was Already There; Vids Avatar Competition Is HeyGen
This AI Mode app-connection update directly responds to the external app integration capabilities already available in OpenAI’s ChatGPT and Anthropic’s Claude. Google Vids’ selfie-avatar feature moves into the existing market for video AI companies such as HeyGen, Synthesia, Captions, and D-ID.
By comparison, OpenAI’s Sora (now discontinued) allowed users to create videos including public figures, such as featuring the likeness of OpenAI CEO Sam Altman. Google Vids’ avatar design builds on this by adding safety measures such as SynthID watermarking, account binding, and age restrictions.
FAQ
In which regions is Google AI Mode app linking currently available, and which apps does it support?
According to Google’s announcement, app linking is currently only available in the United States. It currently supports Instacart, Canva, and YouTube Music; Google says it will soon expand support for more apps.
What safety restrictions does Google Vids selfie avatar feature have?
According to Google’s announcement, each selfie avatar includes an invisible SynthID watermark and is linked to the account holder’s Google account; in certain regions, access is limited to users aged 18 and above.
Which Gemini model does Google Vids integrate, and what features does it have?
According to Google’s announcement, Vids integrates Gemini Omni (Google’s multimodal model). It supports generating videos from text prompts and image references, as well as background replacement, brightness adjustments, effect additions, and newly added step-by-step video editing features.
