Speekit

Easier, faster, and more accessible for everyone. One click and just speak it.

Comment 3

Inspiration

The web is full of information and opportunities but for millions of people, it’s still not truly accessible.
Visually impaired or elderly users often struggle to read small text, navigate complex layouts, or find the right information quickly.
We wanted to create something that would make the web more human, by allowing users to simply ask questions and listen to answers instead of struggling with the interface.
That idea became Speekit, a Chrome extension that uses AI to make browsing accessible, conversational, and effortless.

What it does

Speekit transforms any website into an interactive experience.
When a user asks a question while browsing, Speekit:

  1. Analyzes the content of the current webpage and identifies the most relevant information related to the query.
  2. Sends this context to Chrome Prompt AI (Gemini), which generates an intelligent, natural-language answer.
  3. Customizes the AI’s personality, users can choose between funny, friendly, casual, or formal speaking styles.
  4. Converts the answer into speech using the Google Cloud Text-to-Speech (TTS) API, offering two male and two female voice options.

The entire process is optimized for speed, ensuring that the response is read out loud almost instantly.
Speekit turns passive reading into active, accessible conversation with the web.

How we built it

We developed Speekit as a Chrome extension using:

  • JavaScript / HTML / CSS for the UI and extension logic
  • Chrome APIs for content extraction and context analysis
  • Gemini AI (Chrome Prompt API) for natural language understanding and response generation
  • Google Cloud Text-to-Speech API for vocalizing the answers
  • TF-IDF Algorithm for relevant information extraction.

Our architecture is fully asynchronous, ensuring smooth, low-latency communication between the LLM and TTS pipeline.
We focused on creating a clean and intuitive interface that requires minimal effort from the user, just one question and one click.

Challenges we ran into

  • Context extraction: Selecting the right parts of the webpage to send to Gemini without overloading the model.
  • Latency optimization: Making sure that AI processing and TTS playback happen quickly enough for real-time interaction.
  • UI accessibility: Designing a minimalist interface that is usable for visually impaired and elderly users with clear fonts, contrast, and voice feedback.
  • Voice naturalness: Balancing processing speed and realistic speech tone across different voice styles.

Each of these challenges required careful iteration, testing, and optimization.

Accomplishments that we're proud of

  • Building a fully functional Chrome extension that seamlessly integrates Gemini AI and Google TTS.
  • Achieving fast response times despite multiple API calls.
  • Designing a minimal and inclusive interface accessible to all kinds of users.
  • Allowing full personality and voice customization, giving users more control over how AI communicates with them.
  • Making the web feel more alive and inclusive — especially for users who are often left behind by traditional design.

What we learned

Through Speekit, we learned how to:

  • Efficiently connect multiple Google AI services within a Chrome extension.
  • How to extract relevant information from a webpage based on a user’s question.
  • Optimize performance for real-time, low-latency AI interactions.
  • Prioritize accessibility as a design foundation, not an afterthought.
  • Understand the impact of personalization on user experience — even small customizations make a big difference.

This project deepened our understanding of how technology can bridge accessibility gaps and make AI feel truly personal.

What's next for Speekit

We’re planning to:

  • Expand multilingual support so users can ask and listen in different languages.
  • Enhance the ability to extract relevant information from any type of website.
  • Improve content summarization for long or complex pages.
  • Integrate real-time translation for global accessibility.
  • Release Speekit on the Chrome Web Store to make it available for everyone.

Our vision is to make Speekit a trusted AI accessibility companion — bringing the power of conversation to the web for all users.

Built With

Share this project:

Updates