deflato 1.3.0

Deflato — compress images & documents, generate and edit AI images

Deflato Python SDK

Compress images, documents, and generate AI images with one line of code. Powered by deflato.com.

Install

pip install deflato

Quick Start

from deflato import Deflato

client = Deflato("your-api-key")

# Compress an image
result = client.compress("photo.jpg", format="WEBP", quality=75)
print(f"Saved {result.savings_percent}% -> {result.output_path}")

# Compress a PDF document
doc = client.compress_document("report.pdf", quality=70)
print(f"PDF: {doc.savings_percent}% smaller")

# Generate an AI image
img = client.generate("A sunset over mountains", model="fast")
print(f"Generated -> {img.output_path}")

# Edit an image with AI
edited = client.edit("Add sunglasses to the person", input_path="portrait.jpg")
print(f"Edited -> {edited.output_path}")

# Get alt-text for an image
desc = client.describe(input_path="photo.jpg")
print(desc.description)

API

Deflato(api_key, base_url="https://deflato.com")

---

client.compress(input_path, output=None, quality=80, max_dimension=0, format="JPEG", strip_exif=True)

Compress an image file.

Returns CompressResult:

• success: bool
• input_size: int (bytes)
• output_size: int (bytes)
• savings_percent: int
• output_path: str
• output_format: str

---

client.info(input_path, quality=80, format="JPEG")

Get compression info without saving the file. Returns CompressResult.

---

client.compress_batch(input_paths, output_dir=None, quality=80, format="JPEG")

Compress multiple files. Returns list of CompressResult.

---

client.compress_document(input_path, output=None, quality=75, max_dimension=1920)

Compress images inside PDF, DOCX, PPTX, or XLSX documents. Returns CompressResult.

---

client.convert_format(input_path, output_format, output=None)

Convert an image to a different format (JPEG, PNG, WEBP, AVIF). Returns CompressResult.

---

client.generate(prompt, output=None, model="photo", aspect_ratio="1:1", style="any", resolution="1K", format="png")

Generate an image from a text prompt using AI.

Param	Default	Description
model	photo	fast, photo, design, or icon
aspect_ratio	1:1	Aspect ratio
style	any	Style for design/icon models
resolution	1K	1K, 2K, or 4K (photo only)
format	png	png or jpg

Returns GenerateResult: success, output_path, output_size, model, prompt

---

client.edit(prompt, input_path=None, image_url=None, output=None, model="fast", format="png")

Edit an existing image using AI instructions (img2img).

Param	Default	Description
input_path	—	Path to local image file
image_url	—	URL of the image (alternative)
model	fast	fast (~10s) or precise (~30s)
format	png	png or jpg

Returns EditResult: success, output_path, output_size, model, prompt

---

client.describe(input_path=None, image_url=None)

Generate AI alt-text / caption for an image. Free — no rate limit.

Returns DescribeResult: success, description

---

Supported Formats

Image input: JPEG, PNG, WEBP, HEIC, TIFF, BMP, GIF, RAW (20+ formats)

Image output: JPEG, PNG, WEBP, AVIF

Document input: PDF, DOCX, PPTX, XLSX

Get API Key

1. Sign up at deflato.com
2. Go to Account -> Generate API Key

License

MIT