upsclr / README.md
areebsatin's picture
Bump Gradio to 5.15.0 to fix bool iteration crash
ea6c96d
|
Raw
History Blame Contribute Delete
2.28 kB
---
title: 4K Upscaler Pro
emoji: ⬆️
colorFrom: purple
colorTo: blue
sdk: gradio
sdk_version: "5.15.0"
python_version: "3.10.13"
app_file: app.py
pinned: false
license: mit
---
# ⬆ 4K Upscaler Pro
AI-powered image & video upscaler supporting up to 4K resolution (3840×2160).
## Features
- 🖼 Image upscaling (PNG, JPG, WEBP)
- 🎬 Video upscaling (MP4, AVI, MOV)
- 🤖 Real-ESRGAN AI upscaling (GPU mode) + Lanczos fallback (CPU)
- 📡 Full REST API for n8n, Zapier, or any automation tool
- 🎯 Targets: 4K, 2K, 1080p, or 2× original
## Deployment Guide
### Step 1: Create Hugging Face Space
1. Go to https://huggingface.co/new-space
2. Name it: `4k-upscaler`
3. SDK: **Gradio**
4. Hardware: **CPU Basic** (free) or **T4 GPU** for Real-ESRGAN
### Step 2: Upload Files
Upload all essential files to your Space:
- `app.py`
- `core.py`
- `requirements.txt`
- `pyproject.toml`
- `packages.txt` (adds ffmpeg for video rendering)
### Step 3: Get Your API URL
Your API will be live at:
```
https://<your-username>-4k-upscaler.hf.space/api/predict
```
---
## n8n Integration
### Image Upscaling via n8n HTTP Request Node
```
Method: POST
URL: https://<your-username>-4k-upscaler.hf.space/api/predict
Headers:
Content-Type: application/json
Authorization: Bearer <YOUR_HF_TOKEN> (only if Space is private)
Body (JSON):
{
"data": [
"data:image/png;base64,<BASE64_IMAGE>",
"Lanczos (Fast)",
"4K (3840×2160)"
]
}
```
### n8n Workflow Steps
1. **Trigger** (e.g., HTTP webhook or file watch)
2. **Read Binary Data** → convert image to base64
3. **HTTP Request** → POST to HF Space API
4. **Extract Result**`{{ $json.data[0] }}` = base64 upscaled image
5. **Write File** → save the upscaled image
### Video Upscaling (fn_index: 1)
```json
{
"fn_index": 1,
"data": [
"<video_url_or_base64>",
"Lanczos (Fast)",
"4K (3840×2160)"
]
}
```
---
## API Reference
| Parameter | Values |
|-----------|--------|
| Method | `Real-ESRGAN (AI)` or `Lanczos (Fast)` |
| Target | `4K (3840×2160)`, `2K (2560×1440)`, `1080p (1920×1080)`, `2x Original` |
## Hardware Recommendation
- **CPU (Free):** Lanczos fast upscaling works perfectly
- **T4 GPU ($0.60/hr):** Enables Real-ESRGAN AI upscaling for dramatically better quality