https://github.com/lordbaldwin1/yt-qna-server
API for uploading YouTube video URL, fetching transcript, chunking transcript, embedding transcript, and semantic search on transcript.
https://github.com/lordbaldwin1/yt-qna-server
docker docker-compose drizzle-orm express pgvector postgresql typescript
Last synced: 3 months ago
JSON representation
API for uploading YouTube video URL, fetching transcript, chunking transcript, embedding transcript, and semantic search on transcript.
- Host: GitHub
- URL: https://github.com/lordbaldwin1/yt-qna-server
- Owner: lordbaldwin1
- License: mit
- Created: 2025-05-11T01:25:59.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2025-05-12T07:56:07.000Z (5 months ago)
- Last Synced: 2025-05-12T21:08:50.754Z (5 months ago)
- Topics: docker, docker-compose, drizzle-orm, express, pgvector, postgresql, typescript
- Language: TypeScript
- Homepage:
- Size: 66.4 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# yt-qna-server
This backend allows users to upload YouTube video URLs, fetches and stores video metadata and transcripts, generates embeddings for transcript chunks, and enables users to ask questions about the video content using semantic search and Gemini AI.
## NOTES FOR WORK
- LOOK INTO AND THINK ABOUT WHAT MY API RESPONSE TYPES SHOULD BE.
- I DONT NEED TO RETURN SUCCESS BECAUSE I CAN ALREADY CHECK RESPONSE.OK
## Features
- Upload YouTube video URLs and store metadata, transcript, and summary
- Chunk and embed transcripts for semantic search
- Ask questions about a video and get AI-generated answers with relevant timestamps
- Store and retrieve Q&A history per video
- List all uploaded videos
## Setup
1. Clone the repository
2. Install dependencies:
```bash
npm install
```
3. Set up your `.env` file with the required environment variables:
- `DATABASE_URL`
- `POSTGRES_USER`
- `POSTGRES_PASSWORD`
- `POSTGRES_DB`
- `YOUTUBE_API_KEY`
- `GEMINI_API_KEY`
4. Start the backend:
```bash
npm run dev
```
5. Ensure your Postgres database is running with the `pgvector` extension enabled.
## Main API Endpoints
### Video Endpoints
- `POST /api/videos` — Upload a YouTube video URL. Fetches metadata, transcript, and stores embeddings.
- `GET /api/list-videos` — List all uploaded videos.
### Q&A Endpoints
- `POST /api/ask-question` — Ask a question about a video. Returns an AI-generated answer and the relevant timestamp.
- `GET /api/list-conversation/:videoId` — List all questions and answers for a specific video.
## Technologies Used
- Node.js, Express
- Drizzle ORM
- PostgreSQL with pgvector
- Gemini AI (Google GenAI)
- YouTube Data API
## License
MIT