File-Forge
File Forge
File Forge is an all-in-one, web-based file manipulation tool. Upload a file, process it in seconds, and download the result — no software installation required.
Live App: https://file-forge.onrender.com
Features
PDF Tools
| Feature | Description |
|—|—|
| Remove Password | Unlock password-protected PDF files instantly |
| Convert to Word (Standard) | Fast PDF → DOCX conversion using pdf2docx |
| Convert to Word (AI Layout Recovery) | High-fidelity conversion using PaddleOCR — preserves tables, columns, and complex layouts |
| Extract Pages | Pull specific pages or page ranges (e.g. 1,3,5-10) out of a PDF into a new file |
Image Tools
| Feature | Description | |—|—| | HEIC → JPEG | Convert Apple HEIC/HEIF photos to universally compatible JPEG | | Resize Image | Resize by exact dimensions, percentage, or target file size (KB) | | Crop Image | Crop images using a visual drag-and-drop cropper (powered by Cropper.js) |
Workflow Builder
Chain multiple operations into a single pipeline. Example: unlock a PDF, then convert it to Word — all in one step with real-time progress streaming via Server-Sent Events (SSE).
Tech Stack
- Backend: Python 3.10+, FastAPI, Uvicorn
- PDF Processing:
pikepdf,pdf2docx,PaddleOCR(PPStructure),PyMuPDF - Image Processing:
Pillow,pillow-heif - Frontend: Vanilla HTML/CSS/JS, Cropper.js
- Deployment: Render, Docker
Project Structure
File-Forge/
├── main.py # FastAPI app — all API routes
├── requirements.txt # Python dependencies
├── Dockerfile # Container build config
├── render.yaml # One-click Render deployment config
├── scripts/
│ ├── pdf_utils.py # PDF password removal, conversion, page extraction
│ ├── image_utils.py # HEIC conversion, resize, crop
│ ├── utils.py # Shared helpers
│ ├── security_utils.py # Input sanitization utilities
│ └── fix_models.py # PaddleOCR ONNX model setup script
├── static/
│ ├── index.html # Single-page frontend
│ ├── script.js # All frontend logic
│ └── style.css # Styles
├── models/ # PaddleOCR ONNX model files (auto-downloaded)
├── uploads/ # Temporary upload staging (auto-cleared)
├── outputs/ # Processed files (auto-deleted after download)
└── tests/
├── conftest.py # Fixtures (mock PDFs, test client setup)
├── test_main.py # Integration tests for API endpoints
├── test_pdf_utils.py # Unit tests for PDF processing logic
├── test_image_utils.py # Unit tests for image processing logic
└── ... # Additional benchmarks and edge-case tests
API Reference
All endpoints accept multipart/form-data. Authentication is via X-API-Key header (only enforced when FILE_FORGE_API_KEY env var is set; omit in local dev).
PDF Endpoints
POST /api/pdf/remove-password
Remove the password from a protected PDF.
| Field | Type | Required | Description |
|---|---|---|---|
file |
File | Yes | The PDF file |
password |
string | Yes | The PDF password |
POST /api/pdf/convert-to-word
Convert a PDF to DOCX format.
| Field | Type | Required | Description |
|---|---|---|---|
file |
File | Yes | The PDF file |
use_ai |
bool | No | Use PaddleOCR for layout-aware conversion (default: false) |
password |
string | No | Password if the PDF is encrypted |
POST /api/pdf/extract-pages
Extract a subset of pages from a PDF.
| Field | Type | Required | Description |
|---|---|---|---|
file |
File | Yes | The PDF file |
pages |
string | Yes | Comma-separated pages or ranges, e.g. 1,3,5-10 |
password |
string | No | Password if the PDF is encrypted |
Image Endpoints
POST /api/image/heic-to-jpeg
Convert a HEIC/HEIF image to JPEG.
| Field | Type | Required | Description |
|---|---|---|---|
file |
File | Yes | The HEIC/HEIF image |
quality |
int | No | JPEG quality 1–95 (default: 95) |
POST /api/image/resize
Resize an image.
| Field | Type | Required | Description |
|---|---|---|---|
file |
File | Yes | The image |
mode |
string | Yes | dimensions, percentage, or target_size |
width |
int | No | Target width (for dimensions mode) |
height |
int | No | Target height (for dimensions mode) |
percentage |
int | No | Scale factor (for percentage mode) |
target_size_kb |
int | No | Target file size in KB (for target_size mode) |
POST /api/image/crop
Crop an image to a rectangle.
| Field | Type | Required | Description |
|---|---|---|---|
file |
File | Yes | The image |
x |
int | Yes | Left edge of crop box |
y |
int | Yes | Top edge of crop box |
width |
int | Yes | Width of crop box |
height |
int | Yes | Height of crop box |
Workflow Endpoint
POST /api/workflow/execute
Execute a multi-step pipeline on a file. Returns a Server-Sent Events (SSE) stream with real-time progress.
| Field | Type | Required | Description |
|---|---|---|---|
file |
File | Yes | The input file |
steps |
JSON string | Yes | Array of step objects (see below) |
Step object format:
[
{ "type": "remove_password", "label": "Unlock PDF", "config": { "password": "secret" } },
{ "type": "pdf_to_word", "label": "Convert to Word", "config": { "use_ai": false } }
]
Available type values: remove_password, pdf_to_word, heic_to_jpeg, resize_image, crop_image.
Download Endpoint
GET /api/download/{filename}
Download a processed file. The file is automatically deleted from the server after download.
Getting Started (Local Development)
Prerequisites
- Python 3.10+
- pip
Installation
git clone https://github.com/BhurkeSiddhesh/File-Forge.git
cd File-Forge
pip install -r requirements.txt
Run the Server
python main.py
The app will be available at http://127.0.0.1:8001.
Environment Variables
| Variable | Description | Default |
|---|---|---|
PORT |
Port the server listens on | 8001 |
FILE_FORGE_API_KEY |
API key for authentication (leave unset to disable auth in dev) | (none) |
Running Tests
This project uses pytest. Run the full test suite with:
python -m pytest
Key test files:
tests/conftest.py— fixtures including auto-generated mock PDFstests/test_main.py— integration tests for all API endpointstests/test_pdf_utils.py— unit tests for PDF processing logictests/test_image_utils.py— unit tests for image processing logic
Tests create and clean up all temporary files automatically.
Deployment
Deploy to Render (Recommended)
- Fork this repository and log in to Render.
- Connect your GitHub repository.
- Render will auto-detect
render.yamland configure the service. - Click Deploy.
[!IMPORTANT] This app uses PaddleOCR for high-quality PDF conversion, which requires at least 1 GB of RAM. The Render Starter plan or higher is required. The Free Tier (512 MB) will crash during AI model initialization.
Deploy with Docker
docker build -t file-forge .
docker run -p 8001:8001 file-forge
GitHub Pages
GitHub Pages is not supported — it only serves static files and cannot run the Python backend. Use Render, Railway, Fly.io, or any platform that supports Docker or Python WSGI/ASGI apps.