A Node.js proof-of-concept for semantic search over transcribed audio content using OpenAI embeddings and Pinecone.
Audio / Video → Whisper → Transcript → Chunking → OpenAI Embeddings → Pinecone → Semantic Query
- Audio transcription with Whisper
- Text chunking for embedding generation
- OpenAI embedding generation
- Vector storage and similarity search with Pinecone
- Node.js-based search workflow
Node.js · OpenAI API · Whisper · Pinecone · Axios · UUID
Create a .env file from .env.example:
OPENAI_API_KEY=your-openai-api-key
PINECONE_API_KEY=your-pinecone-api-keyNever commit credentials or API keys.
npm install
node index.js├── embeddings.js
├── pinecone.js
├── query.js
├── whisper.js
├── index.js
└── .env.example
Proof-of-concept / learning project.