Manifest V3 Chrome Extension in Development

Navigate the web with natural voice interaction.

An open-source Chrome extension transforming web browsing for visually impaired users. Featuring powerful AI screen analysis, comprehensive voice control, and seamless switching between cloud and privacy-focused local AI.

Watch Demo

Redefining Web Accessibility

Through intuitive AI chat, powerful screen analysis, and comprehensive voice control, EchoNav AI transforms how you interact with the web.

AI Screen Analysis

Instantly understand webpage content. Analyze screenshots or full-page text with Gemini or your local LLMs to get comprehensive context.

Full Voice Control

Use your voice to navigate, ask questions, and control the browser completely hands-free. Includes precise Voice-to-Text and Text-to-Speech.

Dual AI Modes

Switch with a single click between powerful Cloud AI (Gemini, Mistral) and privacy-focused Local AI (Ollama, LM Studio, vLLM).

Web Page Interaction

Tell the AI to click buttons or type text for you (e.g., 'click the login button', 'type hello world in the search bar').

Cross-Platform Support

Works seamlessly on any Windows, macOS, or Linux computer running Google Chrome. Built securely on Manifest V3 architecture.

Privacy First

No data collection. All settings stored locally. With Local AI mode, your screen data never leaves your computer.

System Architecture

A robust, scalable integration between the client browser, built-in device models, and remote servers.

User Layer

Visually Impaired

Audio Feedback via Speaker

Motor Disability

Voice Commands via Mic

Client-Side (Chrome Browser)

Gemini Nano Built-in Chrome API
Content Scraping JS Text & Image Extraction
Core Extension
Text to Speech
Speech to Text
Intent Classification
Action Executor
Step-by-step Planner

Server-Side Services

Managed LLMs

Public API Calls

Google Gemini Mistral AI
Self-Hosted LLMs

Local API Calls (Privacy)

Ollama LM Studio vLLM Gemma 3
GCP Infrastructure

API Gateway → Cloud Functions → Vertex AI

Proving Our Impact

Accessibility must be measurable. We are conducting rigorous beta testing across diverse user groups to validate success and ensure true digital independence.

500+
Beta Testers

Active participation targeted for Phase 2, spanning diverse visual impairment demographics.

-40%
Task Time

Target reduction in time required to complete complex web forms and checkouts.

-60%
Navigation Errors

Expected decrease in user dead-ends and accidental misclicks via safe action execution.

95%
Satisfaction Score

Target User Satisfaction (CSAT) rating based on qualitative post-task surveys.

A Collaborative Ecosystem

Building the future of web navigation requires bringing together key stakeholders across the accessibility spectrum.

End Users & Advocacy Groups

Blind and low-vision individuals driving our feature roadmap through continuous feedback.

Open Source Community

Developers and AI researchers contributing to the core A-Eye engine and local LLM integrations.

Enterprise & Government Partners

Institutions requiring reliable, WCAG-compliant solutions to serve all citizens safely.

Seeking Endorsements

We are actively seeking pilot partnerships, grants, and official endorsements from leading organizations to validate and scale our technology:

National Blind Advocacy Orgs
Web Accessibility Initiatives (W3C)
Gov. Disability Departments
Tech & AI Ethics Foundations
Become a Partner Organization

Configuration: Setting Up Your AI

You must configure at least one AI mode to enable the extension. Switch seamlessly between modes depending on your privacy and power requirements.

Local AI Setup

Ideal for privacy. All processing happens on your computer.

Option A: Ollama

  • Install Ollama on your system.
  • Set CORS: Run setx OLLAMA_ORIGINS "chrome-extension://*" /M (Windows)
  • Run a model: ollama run gemma3:4b (or 12b/27b based on VRAM)
  • In extension settings, ensure "Local Model Name" matches.

Option B: LM Studio

  • Search and download google/gemma-3 in LM Studio.
  • Go to 'Local Server' tab and click 'Start Server'.
  • Match "LM Studio Model Name" in extension settings.

Option C: Gemini Nano (Chrome Built-in)

  • Requires Chrome Canary. Open chrome://flags.
  • Enable "Prompt API for Gemini Nano" & "Enables optimization guide".
  • Trigger download via console: await LanguageModel.create().

Cloud AI Setup

Ideal for maximum power, speed, and performance.

Option A: Google AI Studio

  • Visit Google AI Studio.
  • Click Get API Key > Create API Key.
  • Copy the key and paste it into the "Cloud API Key" field in the extension's Settings.

Option B: Mistral AI

  • Visit Mistral AI Platform and log in.
  • Navigate to 'API Keys' and create a new key.
  • Paste it into the "Mistral API Key" field in settings.

Advanced: Google Cloud (Vertex AI)

For advanced users who want to manage their own GCP infrastructure, we support routing via custom API Gateways to Vertex AI. See repository docs for details.

Commands & Shortcuts

Action Shortcut Voice Command Example
Open Side Panel Alt+Shift+Q -
Toggle AI Mode Alt+Shift+1 -
Toggle Voice Input Alt+Shift+2 -
Repeat Response Alt+Shift+3 -
Capture Visible Area - "Take a screenshot"
Capture Full Page - "Take a scrolling screenshot"
Analyze Text Content - "Analyze content"

Privacy as a Priority

EchoNav AI (powered by the A-Eye core) is built from the ground up to respect your data. Web accessibility should never compromise your security.

No Data Collection

The extension does not collect or transmit any personal data, telemetry, or browsing history to our servers.

Secure Local Storage

Your settings and API keys are stored solely in your browser's local storage, completely isolated and inaccessible by external websites.

True Local AI Operation

When utilizing Local Mode (Ollama, LM Studio, or Gemini Nano), all data extraction and AI processing remain entirely on your physical machine. Zero cloud pings.

Partner With Us

Interested in partnering with EchoNav AI? Fill out the form below and our team will get back to you.