MeetingMind AI

• 

  MeetingMind AI Dashboard — Upload audio, get instant transcripts, AI summaries & action items

• 

  6 Core Features — Smart Transcription, Speaker ID, AI Summaries, Action Items, Ask AI & Export

• 

  Architecture — AssemblyAI Speech-to-Text & LeMUR powering the full voice intelligence pipeline

💡 Inspiration

Professionals spend countless hours every week sitting in meetings, taking manual notes, and trying to recall who agreed to which action item. MeetingMind AI was built to automate meeting intelligence entirely — transforming raw audio recordings into structured summaries, speaker timelines, action items, and an interactive Q&A assistant.

⚙️ What it does

• Smart Speech-to-Text: High-accuracy transcription with automated punctuation and formatting.
• Speaker Diarization: Identifies who said what across multiple speakers with clear timelines.
• AI Executive Summaries: Condenses lengthy conversations into clear bullet points and key takeaways using AssemblyAI LeMUR.
• Automated Action Items: Extracts to-do lists, deliverables, and assignments starting with actionable verbs.
• Interactive "Ask AI": Allows users to chat directly with their transcript to clarify details, draft follow-up emails, or search topics.
• Multi-Input Support: Upload audio files (.mp3, .wav, .m4a), paste public audio URLs, or record live audio in the browser.

🛠️ How we built it

• AI & Speech Layer: AssemblyAI Speech-to-Text API for transcription & speaker labeling, and AssemblyAI LeMUR for LLM reasoning and Q&A.
• Backend: Node.js and Express for file handling and API orchestration.
• Frontend: Vanilla HTML5, CSS3 with a glassmorphism dark theme, dynamic waveform visualizer, and responsive UI.

🚧 Challenges we ran into

• Orchestrating multi-speaker audio processing seamlessly in real-time.
• Prompting LLM models (LeMUR) to extract clean, deduplicated action items from spoken conversation.
• Handling audio streaming and large file uploads efficiently.

🏆 Accomplishments that we're proud of

• Delivering an end-to-end working voice intelligence dashboard with sub-second response times.
• Creating an intuitive and aesthetically modern UI with zero external bulky frontend framework dependencies.

📚 What we learned

• Deep integration patterns with modern speech-to-text models and voice intelligence pipelines.
• How to structure audio transcripts for optimal retrieval and LLM context extraction.

🚀 What's next for MeetingMind AI

• Real-time live meeting transcription via WebSockets.
• Direct Zoom, Google Meet, and Microsoft Teams calendar bots.
• Automated email digests sent to all meeting participants.

Built With

• assemblyai
• css3
• express.js
• html5
• javascript
• lemur
• node.js
• speech-to-text