https://github.com/berkayildi/feedshot
YouTube video analysis + comic-book infographic generator.
https://github.com/berkayildi/feedshot
ai anthropic fastapi gemini infographics mcp python react typescript vercel youtube
Last synced: 29 days ago
JSON representation
YouTube video analysis + comic-book infographic generator.
- Host: GitHub
- URL: https://github.com/berkayildi/feedshot
- Owner: berkayildi
- License: mit
- Created: 2026-03-31T20:37:46.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2026-04-10T22:54:34.000Z (3 months ago)
- Last Synced: 2026-04-11T00:23:52.834Z (3 months ago)
- Topics: ai, anthropic, fastapi, gemini, infographics, mcp, python, react, typescript, vercel, youtube
- Language: TypeScript
- Homepage: https://feedshot.vercel.app/
- Size: 15.4 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE
Awesome Lists containing this project
README
# FeedShot
[](LICENSE)
Paste a YouTube link or enter X accounts, provide your API keys, get structured content analysis + a comic-book style infographic. Zero sign-up, stateless, BYO keys.
## Architecture
```
Browser (React/Vite/TypeScript)
|
+---> POST /api/analyze
| |-> Supadata API (transcript extraction, user's key)
| |-> YouTube oEmbed API (video title)
| |-> Anthropic Claude analysis
| |-> Returns: structured JSON (TLDR, takeaways, social hook)
|
+---> POST /api/analyze-x-feed
| |-> X API v2 (fetch posts, user's bearer token)
| |-> Anthropic Claude digest synthesis
| |-> Returns: structured JSON (TLDR, takeaways, notable posts, social hook)
|
+---> POST /api/generate-image
|-> Google Gemini image generation
|-> Returns: base64 PNG infographic
```
## How It Works
**YouTube Mode**
1. Paste a YouTube URL
2. Enter your API keys (Supadata + Anthropic, optionally Gemini)
3. Get structured analysis + optional infographic
**X Feed Mode**
1. Enter X accounts to follow (comma-separated)
2. Enter your API keys (X Bearer Token + Anthropic, optionally Gemini)
3. Get a synthesised digest with key takeaways and notable posts
## Local Development
### Backend
```bash
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn api.index:app --reload --port 8000
```
### Frontend
```bash
cd frontend
npm install
npm run dev
```
The Vite dev server proxies `/api` requests to `localhost:8000`.
### Both (parallel)
```bash
make run
```
### Tests
```bash
pip install pytest
pytest
```
## Stack
| Layer | Technology |
| ---------------- | ---------------------------------------------- |
| Frontend | React 19, Vite, TypeScript, Tailwind CSS v4 |
| Backend | FastAPI, Python 3.12+ |
| AI Analysis | Anthropic Claude (claude-sonnet-4-20250514) |
| Image Generation | Google Gemini (gemini-3.1-flash-image-preview) |
| X Feed | X API v2 (user's bearer token) |
| Transcript | [Supadata](https://supadata.ai) API |
| Hosting | Vercel (static + serverless Python) |
## Related
- [mcp-content-pipeline](https://github.com/berkayildi/mcp-content-pipeline) -- the MCP server this app wraps
## License
MIT