interfaze_ocr_viewer

KO | EN

gitlite — search

#ai #typescript #golang #android #machine-learning #rust #macos #nodejs #java #windows #awesome #cli #openai #open-source

interfaze_ocr_viewer

(★ 21)

A Web Viewer To Visualize OCR Results Produced By https://interfaze.ai

interfaze_ocr_viewer Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# OCR Viewer

A browser-based viewer for overlaying OCR data on top of PDF documents. Built specifically to visualize OCR output produced by **[interfaze.ai](https://interfaze.ai/)**.

https://github.com/user-attachments/assets/f4356d2e-a9a3-4141-b5c5-735d42f60757

## What it does

- Renders a PDF page-by-page in the browser
- Overlays bounding boxes from the OCR JSON on top of the matching PDF page
- Color-codes boxes by confidence level (green / yellow / red)
- Shows the recognized text and confidence score on hover
- Supports switching between **line-level** and **word-level** boxes
- Adjustable zoom from 50% to 300%

## Requirements

- Python 3 (for the local server) — already installed on macOS
- A modern browser (Chrome, Firefox, Safari, Edge)
- A PDF file and its corresponding OCR JSON file exported from [interfaze.ai](https://interfaze.ai/)

## How to run

1. **Clone or download** this project and place your files in the same folder.

2. **Start a local HTTP server** from inside the project folder:

```bash
   cd /path/to/ocr_samples
   python3 -m http.server 8765
   ```

3. **Open your browser** and go to:

```
   http://localhost:8765
   ```

4. **Upload your files** using the drag-and-drop zone or the file pickers:
   - Select your **PDF** file
   - Select the **OCR JSON** file that corresponds to that PDF

5. Click **"View document"** to open the viewer.

> Use the **"Open files"** button in the toolbar at any time to load a different document.

## JSON format

The viewer expects OCR JSON in the format produced by [interfaze.ai](https://interfaze.ai/). It supports both the full wrapper response object and a bare OCR result object with the following structure:

```json
{
  "sections": [
    {
      "text": "...",
      "lines": [
        {
          "text": "line text",
          "bounds": {
            "top_left": { "x": 0, "y": 0 },
            "width": 200,
            "height": 30
          },
          "average_confidence": 0.85,
          "words": [
            {
              "text": "word",
              "bounds": { "top_left": { "x": 0, "y": 0 }, "width": 80, "height": 30 },
              "confidence": 0.91
            }
          ]
        }
      ]
    }
  ],
  "width": 1249,
  "height": 7932,
  "total_pages": 5
}
```

Each entry in `sections` corresponds to one page of the PDF (section 0 → page 1, etc.).

## Confidence colour coding

| Colour | Meaning |
|--------|---------|
| 🟢 Green | Confidence ≥ 70% |
| 🟡 Yellow | Confidence 40–70% |
| 🔴 Red | Confidence < 40% |
# interfaze_ocr_viewer

interfaze_ocr_viewer

interfaze_ocr_viewer Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

interfaze_ocr_viewer

interfaze_ocr_viewer Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README