French Middle PDF to Text.

Use the page controls to reach the page you need. Draw, move or resize a rectangle around a passage, or use the whole page. Keep complete lines and punctuation inside the selection.

You can prepare the current page with the image enhancement controls before recognition. Click Start and complete verification when ready. A different page or region requires a separate recognition request.

Use an unlocked copy of a protected PDF. Large or complex documents can reach memory or processing limits even when they are within the file-size and source-page limits.

The recognition-language picker describes the words in your source. It is independent of the website interface language. Selecting another recognition language opens its dedicated OCR page.

Choose the language that best matches the passage you want to read. A language selection is not a promise that every word, mixed-language line or historical typeface will be recognized correctly.

For Classic, choose a column setting that matches the selected content. A focused single-column region can be easier to review than several unrelated text blocks.

Both modes provide editable text and Word output. Classic can also produce a Searchable PDF retaining the selected image area with its recognized text layer. Future's Simple PDF creates a new text layout rather than retaining the original scan design.

Review the source beside the recognized text regardless of the output format. Neither mode guarantees a perfect transcript or an editable reconstruction of the original tables, pictures and typography.

Zoom the source to inspect names, dates, amounts, identifiers and punctuation. Correct the result directly when a letter is wrong or a line break interrupts a sentence.

TXT and Word downloads use the current text. Future's text-only PDF can use the corrected text as well. Later edits do not rewrite the original OCR layer in a Classic Searchable PDF; keep that distinction in mind when choosing a format.

Advanced editor provides headings, emphasis, lists, alignment, text direction, font choices, colors and highlighting. Simple editor is useful for plain-text corrections. Editing the plain text can remove rich formatting, so choose the workflow that matches your intended output.

Summarize text, Paraphrase text, Fix grammar and Suggest a title are optional server-processed actions. Select a passage for a focused request or leave nothing selected to use the whole editor. A request is limited to 600 words and 12,000 characters; shorten the passage when needed.

Remove accents and vowel marks, normalize number shapes, and use Undo, Redo or Restore to compare changes. These tools are useful for cleanup, but changing spelling or wording is different from transcribing the source. Review any change that could alter meaning.

Use TXT for plain text, or the main Word download for an editable document in a new text layout. Choose PDF with the mode distinction in mind: Classic retains the recognized image area, while Future produces text in a new layout.

For headings, lists and other Smart edit formatting, download Word, PDF or HTML directly from that editor. Apply Changes sends plain text back to the main OCR result; it does not turn the original scanned page into editable layout objects.

Download anything you need before reloading the page or starting another file. A working editor is not a saved project, and exports should be opened and checked in the application where you will use them.

Correct recognition mistakes first, then open Translate text and choose the target language. The request translates the entire current text in the main result; highlighting a passage there does not limit translation to that selection.

Copy translation copies the translated version without replacing the main result. Use translated text replaces the editable OCR text with the translation. Keep a copy of the original transcript if you need both versions.

Translation can change nuance or misread an ambiguous phrase. Review names, quantities and important instructions, and consult the source when the result is uncertain.