What is AI Eyes?
AI Eyes is a local-first, open-source prototype designed to manage how AI agents interact with your screen and audio sessions. It provides a permission-based interface that allows users to invite named agents into their workspace while maintaining full control over sensing states and privacy boundaries.
- Best For: Developers and researchers exploring human-AI co-presence and privacy-first interaction models.
- Pricing: Open-source (Free).
- Category: AI Productivity Tools
- Free Option: Yes ✅
The Problem AI Eyes Solves
Modern AI tools often operate as "black boxes," scraping data or monitoring activity without clear user consent or visible status indicators. This lack of transparency creates significant privacy concerns for professionals who need to share their screen or audio with digital assistants. Users often feel they have no granular control over what an agent can see or hear at any given moment.
Developers and researchers working on human-AI collaboration suffer from the absence of standardized, permission-first interfaces. Without a clear way to toggle agent presence, the interaction feels intrusive rather than helpful. AI Eyes addresses this by building a dedicated interface for co-presence, where the user explicitly manages the agent's access to system resources.
By focusing on the "co-presence" aspect, AI Eyes ensures that the AI agent's state—whether it is observing, reacting, or explaining—is always visible to the human user. In this tutorial, you'll learn exactly how to use AI Eyes — step by step.
How to Get Started with AI Eyes in 5 Minutes
- Visit the official GitHub repository for AI Eyes to access the source code and documentation.
- Review the
docs/INSTALL.mdfile to understand the build requirements for your specific operating system. - Clone the repository to your local machine and install the necessary dependencies for the React-based interface.
- Launch the application in your browser to initialize the local-first environment.
- Grant the browser permission to access your screen, window, or audio inputs to begin your first co-presence session.
How to Use AI Eyes: Complete Tutorial
Step 1: Configuring Session Permissions
The core of AI Eyes is its permission-first architecture. When you start a session, the application will prompt you to select which screen, window, or audio source you wish to share. This ensures that no data is captured without your explicit, intentional selection. Once selected, the application establishes a local connection that remains within your browser environment.
Step 2: Managing Agent Presence
Once your session is active, you can invite specific agents, such as "Codi" or "Lyra," into your workspace. These agents act as placeholders for future AI integrations, allowing you to test the interaction model. You can toggle their presence on or off, which helps in managing the cognitive load of having an AI "present" in your session.
Step 3: Switching Sensing Modes
AI Eyes provides three distinct sensing modes: Quiet, React, and Explain. The "Quiet" mode keeps the agent in a passive state, while "React" and "Explain" modes change how the agent interacts with the incoming data stream. By switching between these modes, you can observe how the UI updates to reflect the agent's current level of engagement.
AI Eyes: Pros & Cons
| Pros | Cons |
|---|---|
| Privacy-focused local-first architecture. | No current AI provider integration. |
| Permission-first design ensures user control. | Lacks semantic audiovisual understanding. |
| Transparent sensing states. | Early prototype stage (v0.1.0). |
| Open-source and accessible. | Requires manual build/install steps. |
AI Eyes Pricing: Free vs Paid
AI Eyes is currently an open-source project, meaning it is entirely free to use and modify. There is no pricing model defined, and the project is maintained as a community-driven prototype. You have full access to the source code and can run it locally without any subscription fees or hidden costs.
Because it is an early-stage prototype, there are no "paid" tiers or premium features. All functionality, including the ability to manage agent presence and sensing modes, is available to anyone who clones the repository. 👉 Check the latest pricing and project updates on the official website.
Who is AI Eyes Best For?
For Developers: This tool is ideal for those interested in building their own AI agents and needing a standardized, permission-based UI to handle screen and audio inputs.
For Researchers: It provides a controlled environment to study human-AI interaction, specifically regarding how users respond to visible sensing states and agent presence.
For Privacy Advocates: It offers a transparent, local-first architecture that ensures data stays on your machine, making it a great starting point for building secure AI-assisted workflows.
Who Should Not Use AI Eyes?
If you are looking for a "plug-and-play" AI assistant that can immediately summarize your meetings or analyze your screen content, AI Eyes is not for you. As of version 0.1.0, it does not connect to any AI provider and lacks the semantic understanding required to perform actual tasks. It is a framework for building, not a finished product.
Additionally, if you are not comfortable with command-line interfaces or building software from source code, the current setup process may be frustrating. The project is intended for those who want to contribute to or experiment with the underlying architecture of co-presence, rather than end-users looking for a consumer-ready application.
Alternatives to AI Eyes
Current alternatives include general-purpose screen recording tools, specialized AI meeting assistants like Otter.ai or Fireflies.ai, and open-source projects focused on multimodal AI agents. However, AI Eyes remains the better choice for those specifically interested in the "co-presence" interaction model and a local-first, permission-centric design that prioritizes user visibility over raw model output.
How We Evaluated AI Eyes
This tutorial was compiled based on the official repository documentation, the project's README, and the stated feature set of the v0.1.0 release. We have focused on the architectural goals and the current capabilities of the interface as described by the creators. This review does not claim to have performed extensive long-term stress testing, as the tool is in an early prototype stage.
Final Verdict: Is AI Eyes Worth It?
AI Eyes is a promising foundation for anyone interested in the future of human-AI co-presence. While it currently lacks functional AI integrations, its commitment to privacy and user-controlled sensing makes it a valuable resource for developers. Score: 7/10.