CogniLama

CogniLama transforms online content into accessible, personalized reading experiences.
Adaptive technology simplifies complex information for diverse cognitive needs and abilities.
Customizable Chrome extension with multiple modes and UI settings for every user preference.
Empowering digital inclusion—making the web clear, focused, and user-friendly for all minds.

Inspiration

Millions worldwide—about 1 in 5—face reading and cognitive challenges like dyslexia, ADHD, and age-related decline. The web’s “one-size-fits-all” design excludes many from fully accessing and engaging with online content. Inspired by the need for true digital inclusivity, CogniLama was born to empower diverse minds by adapting web content to their unique ways of understanding and learning

What it does

CogniLama is an open-source Chrome extension that uses an on-device or hosted large language model (Llama 4 Maverick) to transform web content in real-time. It offers multiple modes—Custom Mode, ELI5 Simplify, Visual Clarity, Mind Illuminator, and Translation, Tech Explainer—each designed to support different cognitive needs. It also provides customizable UI settings such as OpenDyslexic font, sepia theme, and adjustable line, letter, and word spacing to maximize comfort and readability while ensuring privacy through local processing.

How we built it

We combined a state-of-the-art open-source language model optimized for cognitive accessibility with seamless Chrome extension technology. The on-device LLM option was prioritized to protect user privacy by processing data locally. We engineered flexible modes to address a variety of needs, alongside UI personalization features that respect individual preferences. Open-source collaboration enabled us to iterate quickly and incorporate community feedback for continuous improvement.

Challenges we ran into

Balancing powerful AI-driven content adaptation while keeping the extension lightweight enough for on-device operation was complex. Ensuring privacy without sacrificing performance required careful model tuning and efficient coding. Designing a user-friendly interface with highly customizable settings presented UX challenges, especially in making features accessible to neurodiverse users. Lastly, making the tool robust across all websites demanded rigorous testing and adaptive content parsing.

Accomplishments that we're proud of

We successfully created a privacy-first tool that dynamically adapts web content for diverse cognitive needs in real time. The integration of multiple modes offers tailored experiences that users can personalize extensively. Launching with the cutting-edge Llama 4 Maverick open-source model and making the tool open source itself exemplifies our commitment to transparency and community-driven digital inclusion. Positive user feedback and engagement from advocates and developers have validated our approach.

What we learned

We learned the critical importance of combining technical innovation with true empathy for users’ lived experiences. Effective digital accessibility is as much about adaptable design and privacy as it is about advanced AI. Collaboration and iterative user testing were invaluable in refining both functionality and usability. Finally, fostering an open-source community not only accelerates development but enriches the project’s impact and sustainability.

What's next for CogniLama

Going forward, we plan to expand CogniLama’s language support and integrate additional AI models to enhance adaptability. We aim to build more robust analytics for personalized recommendations, refine UI customization further, and optimize performance for mobile browsers. Partnerships with educational and accessibility organizations will help scale adoption. Our vision is to make the web universally accessible, giving every mind the power to engage meaningfully and independently online.