Inspiration

Many people feel nervous or judged when speaking with interviewers, professors, or even peers, especially those from different backgrounds or age groups. TalkItOut is inspired by the idea that practicing with a non-judgmental AI coach can help anyone build confidence, master appropriate slang, understand diverse communication styles, and learn to connect authentically with people from all walks of life.

What it does

TalkItOut creates a judgment-free environment for practicing real conversations.

  • Captures speech with a microphone.
  • Converts audio to text and analyzes tone, slang, and clarity using Gemini AI.
  • Offers real-time, empathetic conversational responses.
  • Stores MP3 recordings and conversation data securely in MongoDB for review and progress tracking.

How we built it:

Frontend: React with TypeScript for building the user interface and conversation dashboard. Backend: Express + Node.js handling API requests, authentication, and data processing. Database: MongoDB storing user data, MP3 files, and conversation analytics. AI Processing: Gemini API for natural language understanding, slang detection, and communication guidance. Speech Capture: Browser-based audio recording translated to text for analysis.

Challenges we ran into

  • Maintaining low latency for real-time voice and text feedback.
  • Efficiently storing and serving large MP3 files from MongoDB.
  • Designing feedback that is consistently constructive and never discouraging.

Accomplishments that we're proud of

  • Built a seamless speech-to-feedback pipeline.
  • Enabled contextual, age and culture-aware slang detection.
  • Created an AI model that is both adaptive and empathetic in tone.
  • Ensured collaborative code through full-stack TypeScript model sharing.

What we learned

  • Language context (age, region, setting) is crucial for authentic conversation analysis.
  • Real-time speech systems demand tight backend optimizations.
  • Empathy in AI coaching requires both robust NLP and sensitive UX design.
  • Gemini’s NLP engine can be tuned for highly customized coaching outcomes.

What's next for TalkItOut

  • Develop dashboards for visualizing progress and patterns.
  • Allow users to customize their AI coach’s style and personality.
Share this project:

Updates