Web Development
Build a Meeting Transcription & Summary Tool
Build an AI meeting assistant — real-time transcription, speaker diarization, summaries, and action items.
INTERMEDIATE
1 students
About This Course
Build a tool that joins your meetings, transcribes everything in real-time, identifies who said what (speaker diarization), generates AI summaries, extracts action items, and sends them to Slack or email when the meeting ends. You'll work with real-time audio processing (WebSockets), speech-to-text (Deepgram/Whisper), speaker identification, LLM summarization (Claude/GPT), and calendar integrations. Built with Node.js, Python (for audio processing), Next.js dashboard, PostgreSQL, and Redis.
INTERMEDIATE Level
28 hours
English
Course Syllabus
10 chapters57 lessons7h 41m
- What we're building — from raw audio to structured meeting notesFree9m
- Architecture overview — audio capture, STT pipeline, diarization, LLM, and storageFree7m
- Choosing your STT — Deepgram vs AssemblyAI vs Whisper (cost, latency, accuracy)Free9m
- Project setup — Node.js backend, Python audio service, Next.js dashboard, PostgreSQLFree7m
- Data model — meetings, transcripts, segments, speakers, summaries, and action itemsFree10m
- Audio fundamentals — sample rates, formats, channels, and encodingFree7m
- Capturing audio from the browser — MediaRecorder and getUserMediaFree6m
- Streaming audio over WebSockets — real-time audio chunks to the serverFree9m
- Audio preprocessing — noise reduction, normalization, and VAD (voice activity detection)Free7m
- Handling different audio sources — microphone, system audio, and meeting bot audioFree9m
- Audio buffering and chunking — batching for efficient STT processingFree7m
Student Reviews
--0 reviews
5
0
4
0
3
0
2
0
1
0
No reviews yet. Be the first to review this course!
Free
Sign in to Enroll
This course includes:
- 57 lessons
- 28 hours of content
- INTERMEDIATE level
- English
- Certificate of completion