Semantic PDF Compression

Extract only what's relevant
from any document

Densify takes a PDF and a query, then returns only the sections that matter — reducing document size by up to 40% while preserving the content you actually need.

Book a Call See How It Works

40%

Size reduction

REST

Simple API

How it works

A single API call. Upload a document, describe what you need, get back only the relevant content.

Upload a PDF and a query

Send any PDF along with a question or task. Densify uses your query to determine what's relevant.

Densify filters semantically

The document is split into chunks and scored by relevance to your query. Only the sections that matter are kept.

Get back only what you need

Receive the compressed text with token counts — significantly smaller, semantically intact.

example.py

import requests

response = requests.post(
    "https://api.densify.dev/densify",
    data={"prompt": "What are the termination clauses?"},
    files={"file": open("contract.pdf", "rb")},
)

result = response.json()
print(result["compressed_text"])
print(f"{result['reduction_percentage']:.0f}% reduction")
# 1,840 tokens → 1,104 tokens (40% reduction)

Let's talk