{
  "slug": "ocr-text-cleaner",
  "name": "gizza-ai/ocr-text-cleaner",
  "version": "0.1.0",
  "title": "OCR Text Cleaner — Fix Ligatures, Confusables & Spacing — gizza.ai",
  "description": "Clean OCR text by expanding ligatures, rejoining hyphenated line breaks, reflowing lines, fixing confusables, and normalizing spacing.",
  "tags": [
    "ocr text cleaner",
    "fix ocr errors",
    "ocr correction",
    "ligature expander",
    "dehyphenate",
    "reflow text",
    "ocr confusables",
    "clean scanned text"
  ],
  "category": "text",
  "urls": {
    "page": "https://gizza.ai/tools/ocr-text-cleaner/",
    "markdown": "https://gizza.ai/tools/ocr-text-cleaner/index.md",
    "descriptor": "https://gizza.ai/tools/ocr-text-cleaner/tool.json",
    "deep_link_example": "https://gizza.ai/tools/ocr-text-cleaner/?text=The%20%EF%AC%81nal%20%20report%20says%20HeIIo%20%2Cworld%20.Next&fix_ligatures=true&join_hyphenated=true&line_breaks=keep&fix_confusables=true&fix_rn=true&fix_spacing=true"
  },
  "cli": "gizza tool ocr-text-cleaner \"The final  report says HeIIo ,world .Next\"",
  "tool": {
    "description": "Deterministically clean the mechanical errors OCR engines leave in extracted text: expand typographic ligatures, rejoin words hyphenated across line breaks, reflow or collapse line breaks, fix letter/number confusables by context, optionally undo the rn->m merge error, and normalize spacing. No dictionary or model — every fix is a fixed rule, so output is reproducible.",
    "parameters": {
      "additionalProperties": false,
      "properties": {
        "fix_confusables": {
          "default": true,
          "description": "Fix letter and number confusion using word-versus-number context (HeIIo becomes Hello, l00 becomes 100, a stray | becomes I); alphanumeric codes like COVID19 are preserved.",
          "type": "boolean"
        },
        "fix_ligatures": {
          "default": true,
          "description": "Expand typographic ligatures (fi fl ff ffi ffl ſt st) into plain letters (fi fl ff ffi ffl ft st).",
          "type": "boolean"
        },
        "fix_rn": {
          "default": false,
          "description": "Aggressively replace the rn/RN to m/M OCR merge error everywhere. Off by default: with no dictionary it also rewrites real words containing rn such as modern.",
          "type": "boolean"
        },
        "fix_spacing": {
          "default": true,
          "description": "Normalize spacing: collapse repeated spaces, remove spaces around punctuation, insert a missing space after sentence and clause punctuation, and trim line ends.",
          "type": "boolean"
        },
        "join_hyphenated": {
          "default": true,
          "description": "Rejoin a word split by a hyphen at a line break into one word; an uppercase continuation keeps the hyphen as a real compound.",
          "type": "boolean"
        },
        "line_breaks": {
          "default": "keep",
          "description": "Line-break handling: keep them, join soft breaks within each paragraph (reflow), or collapse every break into one line.",
          "enum": [
            "keep",
            "paragraphs",
            "all"
          ],
          "type": "string"
        },
        "text": {
          "description": "OCR-extracted text to clean.",
          "type": "string"
        }
      },
      "required": [
        "text"
      ],
      "type": "object"
    }
  }
}