Different artificial intelligence (AI) models, which can create images, answer questions, or code, can now be run locally on Android smartphones. Last week, Google launched a new app that enables users to run various AI models from the Hugging Face platform directly on their phones, without an internet connection.
The app, named Google AI Edge Gallery, is currently available for Android, and an iOS version is expected soon. It searches, downloads, and runs AI models on the phone, using its built-in processor.
AI models that run online in the cloud are usually more powerful, but there could be certain drawbacks. For example, users may be uncomfortable sharing personal or sensitive data with remote servers, or they might want to use AI features when they don’t have Wi-Fi or cellular data.
Google is calling this an “experimental Alpha release”. The app can be downloaded from GitHub.
Google AI Edge Gallery's home screen shows shortcuts to different AI tools like “Ask Image” and “AI Chat.” When you tap on one, you’ll see a list of models that work for that task, including Google’s own Gemma 3n model.
The app also includes a feature called “Prompt Lab,” where users can try out single-task prompts — like summarising or rewriting text. It includes task templates and settings that let users customise how the models behave.
The caveats related to AI operations remain. Google notes that how well the app works depends on the phone. Newer, more powerful phones will run the models faster. Also, larger models take longer to finish tasks, like answering questions about images, compared to smaller models.
Google has invited feedback on the app from developers. It has been released under the Apache 2.0 license, which means it is available for public use , and even for commercial purposes.
The app, named Google AI Edge Gallery, is currently available for Android, and an iOS version is expected soon. It searches, downloads, and runs AI models on the phone, using its built-in processor.
AI models that run online in the cloud are usually more powerful, but there could be certain drawbacks. For example, users may be uncomfortable sharing personal or sensitive data with remote servers, or they might want to use AI features when they don’t have Wi-Fi or cellular data.
Google is calling this an “experimental Alpha release”. The app can be downloaded from GitHub.
Google AI Edge Gallery's home screen shows shortcuts to different AI tools like “Ask Image” and “AI Chat.” When you tap on one, you’ll see a list of models that work for that task, including Google’s own Gemma 3n model.
The app also includes a feature called “Prompt Lab,” where users can try out single-task prompts — like summarising or rewriting text. It includes task templates and settings that let users customise how the models behave.
The caveats related to AI operations remain. Google notes that how well the app works depends on the phone. Newer, more powerful phones will run the models faster. Also, larger models take longer to finish tasks, like answering questions about images, compared to smaller models.
Google has invited feedback on the app from developers. It has been released under the Apache 2.0 license, which means it is available for public use , and even for commercial purposes.
You may also like
South Africa keeper-batter Heinrich Klaasen announces retirement from international cricket
'Different sets of rules for Hindus and Muslims in Bengal': BJP on arrest of Instagram influencer Sharmishtha Panoli
BJP targets Congress as Kharge's photo left out of Goa Statehood Day poster
Mumbai COVID-19 Update: 43-Year-Old Man Passes Away In Thane; Maharashtra Reports 8 Deaths, 500 Active Cases
Combination Loan: What is a combination loan? Banks themselves do not give information about it while taking a loan.