ExpressAI Adds Google’s Gemma 4 31B Model, Bringing Video Analysis to Pro Users
Published
Key Takeaways
- ExpressAI video analysis: Gemma 4 31B enables video uploads, queries, summaries, and content discovery.
- Long-context processing: 256,000-token window helps analyze lengthy videos, documents, conversations, and codebases.
- Privacy-focused rollout: Video analysis uses ExpressAI's secure enclave architecture without data retention or training.
ExpressAI has expanded its AI model lineup with the addition of Gemma 4 31B, a new model developed by Google DeepMind. The update introduces video analysis capabilities to the platform, allowing users to upload videos and interact with them using natural language prompts.
Available immediately for both new and existing Pro subscribers at no additional cost, Gemma becomes the sixth AI model on ExpressAI and the first on the platform capable of understanding text, images, and video within a single model.
As video content continues to grow across workplaces and educational environments, the new integration aims to help users extract information from recordings without having to manually review entire files.
ExpressAI Now Supports Video-Based Queries
Before this update, ExpressVPN's ExpressAI users could analyze documents and images. With Gemma 4 31B, video files are now part of the same workflow.
Users can upload supported video formats, including MP4, MOV, and WebM, through a drag-and-drop interface and ask questions about the content in plain language.
Examples of supported queries include:
- Summarizing a video clip
- Identifying events that occur during specific time periods
- Counting the number of people appearing in a scene
- Listing products shown in a recording
- Determining whether someone entered or left a room
The feature is designed to help users work with a variety of recorded content, including customer interviews, lectures, product demonstrations, meeting recordings, screen captures, and marketing videos.
Instead of manually searching through footage, users can request summaries, identify key moments, review decisions made during meetings, or extract themes from longer recordings.
Large Context Window Supports Longer Files
One of Gemma's key capabilities is its 256,000-token context window, which allows the model to process and retain information across larger files and longer conversations.
For video analysis, the expanded context helps the model understand events throughout extended recordings rather than evaluating isolated segments. The same capability also benefits document analysis, enabling work across lengthy contracts, books, research papers, reports, and large code repositories.
This broader context allows users to ask questions that depend on information spread across an entire file. Examples include comparing statements made at different points in a meeting, identifying recurring customer concerns during calls, following themes throughout a lecture, or connecting sections within a contract.
Expanded Support for Visual and Complex Analysis
While video support is the headline addition, Gemma also strengthens ExpressAI's capabilities across other content types.
The model can analyze documents, screenshots, charts, diagrams, and other visual materials. It also includes a reasoning mode intended for more complex tasks such as:
- Comparing contracts
- Debugging software code
- Evaluating financial models
- Following cause-and-effect relationships in research materials
According to ExpressAI, these capabilities are designed to help users work with both visual and long-form content from a single interface.
Potential Use Cases Across Different Teams
ExpressAI highlighted several areas where Gemma's capabilities could be useful.
Content and media teams can use the model to summarize videos, generate captions, identify highlights, and create searchable video archives.
In education and research environments, Gemma can help organize recorded lectures, generate study materials, summarize lengthy documents, and analyze information across multiple sources within the same conversation.
Marketing teams may use the model to review campaign footage, identify recurring visual elements, and repurpose long-form content into shorter formats.
Customer insight and research teams can analyze user interviews, customer calls, support recordings, and screen captures to identify common themes and significant moments.
Privacy Protections Remain in Place
ExpressAI stated that all interactions involving Gemma are protected by the platform's existing secure-enclave architecture.
According to the company, prompts, uploaded files, and video content are decrypted only within the secure environment. The company also said that user data is not retained, reviewed, or used for model training.
The privacy framework applies to video analysis in the same way it does to other content processed through ExpressAI.
Availability
Gemma 4 31B is now available to all new and existing ExpressAI Pro plan users without any additional charges.
To access the model, users can open their ExpressAI workspace, select Gemma from the model selection menu, and start a new conversation. Video files can be uploaded through the same drag-and-drop process currently used for documents and images.
Related
Get expert insights on threats, breaches, scams, and security trends — delivered every Monday.
Most Popular
Get expert insights on threats, breaches, scams, and security trends — delivered every Monday.
Most Popular