About UGLYPEAR DATA

Who We Are

High-performance document compression × RAG data engineering

RAG DATA PIPELINE

Parse → Clean → Mask → Compress → Chunk → Evaluate: turning enterprise documents into knowledge your AI can understand, trust, and use securely

Guangzhou UGLYPEAR DATA Technology Co., Ltd. is a software company focused on file compression and AI data engineering. The quality of an AI's answers is only as good as the data you feed it. Built on our in-house Rust compression engine and the RAG DATA PIPELINE, we help enterprises turn scattered contracts, reports, drawings and scans into clean, secure, traceable, searchable knowledge — knowledge that performs reliably with any LLM. Everything runs fully on-premises, so data never leaves your network.

We take an engineering-first approach: every capability we claim can be verified in the source code or in the product. No exaggeration, no vaporware.

What We Do

From document compression to RAG data engineering — one complete pipeline

RAG Data Pipeline

A pipeline that runs automatically the moment you drop a file in:

Parse — six formats (PDF/OFD/Word/Excel/PPT/XPS) plus OCR for scans
Clean — strips headers, footers, page numbers and watermarks; deduplicates content
Mask — AI detects and redacts 13 types of sensitive data, with auditable logs
Chunk — smart chunking by document type, with parent-child linking and four-level content enrichment
Evaluate — three-layer quality checks (structure/relation/content), with failures traced to the exact stage

Explore the RAG Data Pipeline →

Eight Industry Solutions

No one-size-fits-all parameters — four-channel processing strategies (text/image/table/structure), tuned per industry:

Eight industries: legal, healthcare, finance, public sector, academia, manufacturing, e-commerce, and general enterprise
100% retention of critical data: case numbers, dosages, financial figures — nothing gets lost
13% better compression on average, 99.2% AI retrieval effectiveness
5–10 ready-to-use presets per industry

Explore industry solutions →

SmartSlim

A compression product family for individuals and enterprises:

Desktop — Tauri 2.0 + React 19 cross-platform app
Server / Network — FastAPI standalone, or Docker + K8s cluster
Rust compression SDK — C ABI, prebuilt for six platforms
11 categories, 80 file formats supported

Explore SmartSlim →

One Company, Three Brands

How the corporate, AI, and product brands fit together

Corporate brand: UGLYPEAR DATA

The company brand — home of the technology foundation and engineering culture

AI brand: UGLYPEAR AI

Our technology brand for the AI era — the RAG Data Pipeline and eight industry solutions

Product brand: SmartSlim

The market-facing product brand — four product editions

Desktop edition

Tauri 2.0 + React 19 cross-platform desktop app, built for individuals

Server / Network editions

Python FastAPI standalone, or Docker + K8s microservices — built for enterprises

Rust compression SDK

C ABI interface, prebuilt for six platforms — made for third-party integration

One engine underneath: all four product editions share the same in-house Rust compression engine, so compression performance, format support, and quality standards stay consistent everywhere.

Engineering You Can Verify

Every capability below is checkable in the source code

In-house Rust compression engine — not a wrapper around open-source tools; from algorithms to scheduling, everything is built in-house
RAG DATA PIPELINE, fully automated — parse → clean → mask → compress → chunk → evaluate; drop a file in and it runs end to end
Three-layer evaluation loop — structure/relation/content checks, with failures traced to the exact chunking, cleaning, or redaction stage
Prebuilt for six platforms — linux-x64/arm64, linux-server-x64/arm64, macos-universal, windows-x64
Tauri 2.0 + React 19 — a modern stack with small installers and low memory footprint; no Electron bloat
Docker / K8s cloud-native architecture — eight services orchestrated with Docker Compose; Kubernetes HPA auto-scaling
C ABI SDK design — a standardized cross-language interface for Python, Node.js, Java, Go, and more
FastAPI async backend — calls the Rust engine natively, pairing fast development with fast execution
Layered security architecture — differentiated protection for desktop and server, keeping every call legitimate
11 categories, 80 formats — images, PDF, OFD, Office, video, audio, text, e-books, and industry & scientific data, fully covered

Where We're Headed

Focused on data quality, sharpening the core pipeline

Continuous Rust Engine Improvement

Better compression ratios and speed
Stronger memory safety and concurrency
Lower resource footprint, more devices supported
Always current with the Rust ecosystem

Deeper RAG Data Pipeline

Broader format coverage in parsing and cleaning
A growing library of industry-specific strategies
Stronger access control and auditing
Better governance of knowledge freshness

More Formats

Wider coverage of specialized file formats
Deeper support for fixed-layout documents like OFD/XPS
Support for emerging media codecs
Prioritized by customer demand

Faster Everywhere

Higher batch-processing throughput
Lower latency on large-file compression
A smoother desktop experience
Better monitoring and observability