> ## Documentation Index
> Fetch the complete documentation index at: https://docs.repli5.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Welcome to the Repli5 platform

> Turn limited real-world photos into balanced, production-ready computer-vision training datasets.

Collecting and labeling enough real-world images—especially rare surface defects, edge-case lighting, or new material finishes—is one of the biggest bottlenecks in industrial computer vision.

**The Repli5 platform** is built for **ML, perception, and quality-inspection teams** who need to close those data gaps without weeks of manual collection or hand-labeling. Starting from your existing part photos, you can synthesize localized physical defects or generate material, color, and lighting variations using natural-language descriptions, reference close-ups, and reusable project templates—complete with automatic COCO segmentation masks, 2D bounding boxes, and automated visual quality evaluation.

***

## Core Workflows & Capabilities

> [card] [Defect Generation](https://docs.repli5.com/workflows/defect-generation.md)
>
> Synthesize realistic, localized surface defects (scratches, dents, tears, wrinkles) onto defect-free (`OK`) part images with automatic COCO masks and 2D bounding boxes.

> [card] [Variation Generator](https://docs.repli5.com/workflows/variation-generator.md)
>
> Transform object colors, surface materials, and lighting in close-up photos while preserving existing geometry, corner wear, and surface defects.

> [card] [Projects & Data Library](https://docs.repli5.com/workspace/projects-and-data-library.md)
>
> Upload `.zip` datasets, organize shared team workspaces with Projects, filter datapoints by custom tags, and export ready-to-train archives.

> [card] [Quality Evaluation & Review](https://docs.repli5.com/evaluation/quality-evaluation.md)
>
> Automatically score every generated image across workflow-specific visual rubrics, review outputs, and regenerate individual images in-place.

***

## Ready to Begin?

> [card] [Quick Start Guide](https://docs.repli5.com/quick-start.md)
>
> Follow the step-by-step walkthrough to upload your first dataset, set up a project, choose the right workflow, and export your augmented training data.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.