echo agent

An AI workspace combining social media content creation and Python code debugging into a single unified dashboard powered by Google Gemini.

Comment

Inspiration

Developers and content creators often juggle multiple platforms—drafting tailored social posts on one end while quick-testing and debugging code on the other. We created EchoAgent to unify social media management and developer utilities into a single, seamless AI workspace powered by Gemini.

What it does

EchoAgent acts as an all-in-one AI assistant workspace featuring two core modules:

  • AI Social Media Creator: Generates optimized posts for platforms like LinkedIn, Twitter/X, and Instagram. Users can customize tones (Professional, Casual & Friendly, Excited & Bold, Informative) and perform iterative refinements on the output.
  • Python Code Corrector: A lightweight browser workspace where users paste raw or broken Python code. Gemini analyzes the script, fixes syntax and logic errors (such as converting pseudocode into runnable turtle graphics), and delivers a detailed breakdown of what went wrong.
  • System Monitoring: Keeps track of real-time workspace health, uptime, and interaction history.

How we built it

  • Frontend: Next.js, HTML/CSS, and JavaScript for an interactive workspace interface.
  • AI Integration: Google Gemini API (@google/genai) for handling both conversational post generation/refinement and deep code syntax error analysis.
  • Backend & Hosting: Python backend APIs deployed on Vercel (https://nn-mu-azure.vercel.app/).

Challenges we ran into

  • Structured Code Analysis: Prompting Gemini to return both clean, runnable Python code and a distinct bug breakdown without breaking the UI layout.
  • Context Preservation: Ensuring tone adjustments and iterative refinements retained the original user context across multiple post-generation requests.

Accomplishments that we're proud of

  • Successfully building an dynamic Python syntax fixer that correctly interprets ambiguous pseudocode (like turning show a square into structured turtle code) and explains the fixes clearly.
  • Designing a clean, responsive single-dashboard UI that seamlessly combines marketing tools with developer utilities.

What we learned

  • Effective prompt engineering strategies for splitting complex tasks (code generation vs. diagnostic feedback).
  • Best practices for handling asynchronous AI state updates in real-time web workspaces.

What's next for EchoAgent

  • Direct Social Publishing: Integrating official APIs to post generated content directly to LinkedIn, Twitter/X, and Instagram.
  • Live Sandbox Execution: Integrating an in-browser WebAssembly/Pyodide runner so fixed Python code can execute directly on the dashboard.
  • RAG-based Context: Allowing users to connect personal documentation or previous posts for even tighter brand voice consistency.
Share this project:

Updates

Submission history