Lexa

Context for your Everyday AI tools

Comment

Inspiration

A lot of time and credits are wasted in GenAI tools due to inefficient prompting and loss of context. Each session starts from scratch, context gets scattered across chats, and users struggle to phrase prompts for optimal results. We built Lexa to fix that — an AI assistant that understands intent, optimizes prompts, and remembers context across tools, making GenAI faster, cheaper, and smarter to use.

What It Does

Lexa acts as your prompt enhancer and context manager, improving how you interact with GenAI tools in two core ways:

  • Prompt Enhancement: It asks clarifying questions and injects relevant context from your past projects to generate precise, effective prompts.
  • Format Optimization: For image/video generation, it automatically structures prompts into JSON with lighting, aesthetics, and technical details for consistently high-quality outputs.

Result: Smarter prompts, persistent context, and reduced token usage — all without sacrificing creativity or control.

How We Built It

  • Frontend: Chrome Extension built with React + Tailwind for an intuitive, Grammarly-like overlay.
  • Backend: Powered by Gemini Nano with a lightweight retrieval layer to store and recall user context.
  • Prompt Engine: Two-step reasoning — (1) clarify and enrich the prompt, (2) output in optimal format (text or JSON).

Challenges

  • Creating a seamless UX that enhances without interrupting the flow.
  • Managing multi-API integrations and ensuring low-latency context retrieval.
  • Balancing automation with user creativity and maintaining schema consistency across GenAI tools.

Accomplishments

  • Built a working Chrome Extension prototype with real-time prompt enhancement.
  • Designed a unified schema for image/video optimization.
  • Developed a universal context layer that dynamically adapts to user projects.
  • Made GenAI prompting feel effortless and human-aware.

What’s Next

  • UI Revamp: Refine the extension’s UX for a smoother, contextual writing experience.
  • Custom Model: Train a lightweight model tailored for prompt enhancement.
  • Team Mode: Enable shared context across collaborators and projects.
  • Prompt Library: Learn from user history to auto-suggest tone, style, and structure.
  • Integrations: Expand to IDEs, Notion, and Figma for a seamless multi-platform experience.

Built With

Share this project:

Updates