ppu-paddle-ocr/cli/help.d.ts

Lightweight, probably the fastest PaddleOCR SDK in TypeScript. Runs anywhere JavaScript runs: Node.js, Bun, Deno, mobile react-native, web browsers, and browser extensions. Docker & CLI supported. The official SDK is browser-only. Accurate text detection

/**
 * Usage text for the `ppu-paddle-ocr` CLI. Printed to stdout for `help` /
 * `--help`, and to stderr (followed by a non-zero exit) on usage errors.
 */
export declare const USAGE = "ppu-paddle-ocr \u2014 PaddleOCR on the command line\n\nUsage:\n  ppu-paddle-ocr <command> [args] [flags]\n\nCommands:\n  recognize <image>      OCR one image (file path or http(s) URL)\n  batch <pattern...>     OCR many images (globs or a list of paths/URLs)\n  stream <pattern...>    OCR many images, printing each result as it finishes\n  download-models        Pre-warm the model cache (~/.cache/ppu-paddle-ocr)\n  clear-cache            Delete the cached model files\n  models                 Print the active models, defaults, and providers\n  help                   Show this help\n  version                Show the installed version\n\nRecognition flags (recognize / batch / stream):\n  --strategy <s>                 per-box | per-line | cross-line  (default per-box)\n  --cross-line-width-factor <n>  bin-pack width multiplier for cross-line (default 1.0)\n  --engine <e>                   opencv | canvas-native  (default opencv)\n  --image-height <n>             recognition input height in px (default 48)\n  --flatten                      flat results in reading order instead of grouped lines\n  --no-cache                     bypass the in-memory result cache\n\nModel overrides:\n  --model <preset>               catalogue preset, e.g. v6-small (default), v6-tiny,\n                                 v5-en-mobile, v5-thai-mobile (see 'models --json')\n  --model-detection <path|url>   override the preset's detection model\n  --model-recognition <path|url> override the preset's recognition model\n  --model-dict <path|url>        override the preset's dictionary\n\nDetection tuning:\n  --max-side-length <n>          longest side before downscale (default 640)\n  --padding-vertical <n>         box padding, fraction of height (default 0.4)\n  --padding-horizontal <n>       box padding, fraction of height (default 0.6)\n  --min-area <n>                 drop boxes smaller than this area in px (default 50)\n  --mean <r,g,b>                 normalization mean (default 0.485,0.456,0.406)\n  --std <r,g,b>                  normalization std dev (default 0.229,0.224,0.225)\n\nSession:\n  --execution-providers <list>   comma-separated, e.g. cpu or cuda,cpu  (default cpu)\n\nBatch / stream:\n  --concurrency <n|auto>         images in flight at once (default auto)\n  --settle                       keep going past a failed image (default on for batch/stream)\n\nOutput:\n  -o, --output <file>            write to a file instead of stdout\n  --json                         emit structured JSON (NDJSON for stream)\n  --pretty                       indent JSON output\n  -q, --quiet                    suppress progress/logs on stderr\n  --verbose                      log each processing step to stderr\n  --debug                        dump intermediate frames to disk\n  --debug-folder <dir>           where --debug writes (default out)\n\nExamples:\n  ppu-paddle-ocr recognize receipt.jpg\n  ppu-paddle-ocr recognize https://example.com/invoice.png --json --pretty\n  ppu-paddle-ocr batch \"scans/*.png\" --strategy cross-line -o results.json --json\n  ppu-paddle-ocr download-models --verbose\n";