An AAC (Augmentative and Alternative Communication) software powered by artificial intelligence, designed to give a voice back to people living with neurodegenerative diseases. Also controllable via eye tracking.
OUR PROJECT
IGOOR is a free, open-source application (AGPLv3-licensed), also controllable via eye tracking, designed to provide people living with neurodegenerative diseases or paralysis with a more fluid and natural way to communicate.
Vous pouvez accélérer son développement en participant au crowdfunding :
PROJECT OVERVIEW
PROJECT OVERVIEW
IGOOR offers sentence prediction, adapts to individual communication styles, and learns from user data to improve accuracy over time.
- Eye-tracking navigation: the software’s intuitive interface can be easily navigated using eye control.
- Text-to-speech with optional voice cloning: allows users to speak with a reconstructed version of their own voice, enhancing emotional connection for both themselves and others.
- Ability to respond to any question asked by a conversation partner (via the device’s microphone), using a mix of default responses, and previously used answers.
- Improved AI prediction accuracy through uploaded documents about the user, describing their biography, family, relationships, current health condition, etc.
- These are used to refine AI responses and can be updated as the condition evolves or new information becomes available.
- Primary needs interface with editable items
- Predictive keyboard: instead of suggesting single words, the keyboard generates complete sentence suggestions based on the first letters typed.
PROJECT HISTORY
The original idea for this application comes from Igor Novitzki, founder of the IGOOR association, who presented it in this Vimeo video in June 2024 (English translation is July 2024).
The video also showcased our very first working POC (proof-of-concept).
Although IGOOR has evolved significantly since this video, and AI technologies have advanced tremendously, all the fundamental principles of the application were already presented and partially implemented, including:
-
Inclusion of the conversation partner;
-
Predictions based on context using all available information sources: conversations and documents provided by the user, weather, emails, etc.
The video was also published on the website you are visiting now.
OUR VALUES
Freedom (as in “free speech”)
This software is free and open-source, meaning it is available online at no cost, along with its source code, for everyone regardless of income.
Access to the source code encourages collaboration among developers, users, and experts, allowing the software to evolve continuously.
Furthermore, to our knowledge, IGOOR is the only software of its kind distributed under the AGPLv3 license, also known as a copyleft license, which prevents a company or startup from taking the project, improving the code internally, and creating a closed proprietary version. Any improvement to the software must remain free and accessible to all, whether it is a desktop application or a cloud service.
This ensures that the benefits of collective work remain dedicated to patients and the community, and cannot be captured in a closed form by private entities in the future.
Hardware agnostic
The tool is designed to operate independently of any specific hardware, whether a microphone, camera, eye-tracking system, or any other input/output device.
Adopting a hardware-neutral approach is a fundamental principle for developing communication aid tools that are truly inclusive and accessible to everyone. By promoting flexibility, economic accessibility, and user freedom of choice, we contribute to creating a more inclusive digital society that respects the diversity of needs.
Initially distributed for Windows, the software is designed to be cross-platform and will be ported to other operating systems in the future (Linux and Mac).
The choice of Windows is due to its compatibility with the most widely available eye-tracking systems on the market.
Offline functionality
We are orienting the development towards a usage that requires minimal Internet connection, which is particularly important for people with disabilities, especially when they are in areas with limited connectivity.
Some software features will always require an Internet connection, such as weather awareness. However, these features are not essential.
Additionally, certain hardware limitations currently prevent us from providing a completely offline solution.
For example, it is technically impossible today to offer IGOOR’s prediction speed without using an external cloud AI provider.
We are confident that these hardware limitations will be overcome quickly, and we are committed to updating our code as soon as possible so that the entire software can operate without the need for external connectivity.
OBJECTIFS
Main goals
- Increased speed to make communication smoother between the user and those around them
- Improved prediction accuracy based on a biographical text document
- Better balance between speed, freedom, and precision of expression
- Speech synthesis integrating the user’s voice, reconstructed through AI from an audio corpus provided by the user
- Simplicity and usability of the user interface
Secondary goals
- Increase the fidelity of the user’s expression style by training on a corpus of texts, images, music, videos, and other user-provided content
- Integrate environmental information to improve response intelligence (GPS location, object detection, weather, local time, scenes, people, and situations via the camera with multimodal AI support)
- Continuously improve the relevance of predicted responses based on accumulated past responses and evolving contextual data
GET INVOLVED
Your contribution is crucial to improve the quality of life for patients with ALS and other conditions affecting communication.
With your donation, hundreds of thousands of patients will be able to better communicate with their loved ones and express their physiological, social, and emotional needs.
You can also get involved in other ways :