Practical user guide
How to use the PicoToolx Image Editor: edit image text, retouch, crop and export
A picture can contain a price, heading, date or label that you need to change. A normal text tool only adds a caption above that picture. This image editor also provides a different workflow: use OCR to locate the existing words, select one of them, repair the pixels behind the old lettering, and draw your replacement in the same area.
The result is an edited image. It is not a recovered Word document, and the original font cannot be identified with certainty from the pixels alone. This guide explains how to get a useful result, when to adjust the controls, and when a busy photograph needs manual retouching.
1. Open a clear image and keep your source copy
Choose Choose image or drag one JPG, PNG or WebP into the upload area. The editor accepts images up to 20 MB, 16 megapixels and 8000 pixels on the longest side. A supported phone photograph may therefore open directly, while a very large scan may need to be resized first. The source file on your device is not overwritten.
For text editing, a sharp, straight image works better than a tiny screenshot enlarged many times. If letters are blurred, cut off, strongly tilted or covered by glare, OCR may miss them. Rotate the image upright and crop unnecessary margins when useful, then run detection. Keep a source copy so you can compare the final result.
2. Detect the text, then click the word you want to change
- Choose English, বাংলা or বাংলা + English according to the writing in your image.
- Choose Individual words for a small correction or Complete lines to replace a whole phrase.
- Click Detect text. The first use downloads the OCR library and language model; later uses may load faster from browser cache.
- Click a highlighted word or line on the canvas. You can also select the recognized text from the Detected text list.
- Change the text in the replacement dialog. Review spelling, size, color and the background preview.
- Click Apply replacement, inspect the image at 100%, then export when ready.
The blue and amber outlines are editing aids. They are not included in your exported picture. A low confidence score means the recognized characters need extra checking; it is not a quality score for your resume, design or document. Verify names, amounts, dates, punctuation and Bengali conjuncts yourself.
What “edit image text” means here: the editor replaces the appearance of existing letters by changing the image pixels in the chosen region. The old lettering is cleaned up and the replacement is rendered as an editable layer during this session. It does not uncover an original font file or turn a flat picture into a fully structured design file.
3. Match the font, color and background carefully
The starting font size and text color are estimates based on the detected region. Choose a different font family when needed, adjust the size, and use bold, italic, underline, line height or letter spacing to improve the match. If you have the matching font file, use Upload matching .ttf / .otf font. Uploading a font improves your available choices; it still does not prove that it was the font used in the original image.
Fit to original box tries to make the replacement fit within the detected text area. A longer phrase may need smaller lettering. Turn fitting off when you want the text to wrap or extend below the original area, then move or resize its layer. Avoid squeezing a long sentence into a tiny label if readability matters.
| Background option | When to use it | What to check |
|---|---|---|
| Sample surrounding color | White documents, plain signs and solid-color graphics | The nearby pixels should represent a clean background, not another letter or a border. |
| Edge blend | Simple, gentle gradients | It interpolates nearby colors; it cannot recreate detailed texture hidden behind text. |
| Chosen solid color | When you know the exact background color | Pick the correct color and check for a visible rectangular patch. |
| Do not remove old pixels | Intentional overlay or adding text over an existing design | The original lettering will remain underneath and may still be visible. |
Use Cleanup padding to include faint edges around the old letters. Too little padding can leave a halo; too much can remove a neighboring character or part of a line. Start small and inspect the replacement. The dialog preview focuses on background repair before the main photo adjustments are applied.
4. Select a text area manually, or remove lettering
If OCR misses a word, choose Select text area and drag a tight rectangle around it. Type the replacement and choose a background repair method. To remove the selected letters without adding new ones, leave the replacement blank and apply the repair. This is useful for a small unwanted label on a plain background.
Manual selection is also useful when a decorative font is readable to you but confusing to OCR. A whole line can be easier to repair than several touching characters. Avoid including unrelated borders, icons or nearby text in the repair area. Large repairs are limited to two megapixels per region to keep browser memory use manageable.
5. Use the clone brush for detailed backgrounds
A grass texture, patterned wall or detailed photograph cannot reliably be rebuilt from a surrounding color. Choose Clone brush, click Set clone source or use Alt+click to select a clean part of the source background, then paint over the unwanted pixels. Choose a source with similar texture and lighting. Short strokes and a small brush give you more control than one large stroke.
The clone brush works on the source background and its repairs. Added text, logos and shapes are separate layers; remove those with Delete layer. You can undo a clone stroke if its texture is wrong. Cloning is manual retouching, not an AI reconstruction of the hidden scene.
Use Select / move to reposition added layers and drag the blue corner to resize them. Image layers keep their proportions unless you hold Shift. Choose Bring forward or Send backward to change the stacking order. The visibility checkbox lets you compare a layer without deleting it. Removing a replacement layer brings its original source lettering back.
6. Crop, resize, rotate and improve the photo
For cropping, choose Crop, drag the selection, and use Apply crop selection in the properties panel. Available ratios include free crop, square, 4:3, 16:9 and the original ratio. For resizing, enter pixel dimensions and keep the aspect-ratio lock on unless you intentionally want to stretch the image. Enlarging cannot recover detail that was missing from the original.
Rotate and flip controls help straighten the overall orientation. Brightness, contrast, saturation, hue, blur, grayscale, sepia and sharpening can be adjusted in Photo adjustments. Presets provide a starting point; use the sliders to refine them. These controls affect the source background and its repairs, while added text and other layers keep their own appearance.
Crop, resize, rotate and flip merge visible edits into image pixels. This keeps the transformation predictable, but those merged text layers are no longer individually editable afterward. Undo restores the earlier state. You can also run OCR again on the transformed picture. Save a project before major changes if you want an additional working copy.
7. Choose the right export format
| Format | Useful for | Important detail |
|---|---|---|
| PNG | Screenshots, text-heavy graphics, logos and transparency | Lossless compression avoids adding JPG artifacts, but files can be larger. |
| JPG | Photographs and broadly compatible uploads | Compression can soften small lettering. Transparency is filled with your chosen background color. |
| WebP | Efficient images for modern websites | Check that the destination accepts WebP. The quality control changes compression. |
Click Export image, choose the format and quality, then download. The export uses the current output dimensions; the screen zoom does not reduce the saved resolution. Detection boxes, crop guides and selection handles are excluded. Open the saved image and review it at its actual size before publishing or sending it.
If you need an additional size reduction afterward, try the Image Compressor. For a simpler dimension-only task, the Image Resizer may be enough. If you only need to copy writing rather than change the picture, use the Image to Text Converter.
8. Save a project when you want to continue editing
A PNG, JPG or WebP is a finished raster image. It does not preserve the editor's separate text and shape layers. Use Save project to download an editable JSON project with the source image, current settings, layers and uploaded fonts. Use Open saved project to continue later. Undo history and OCR detection results are not included; re-run detection when needed.
The project contains the original source image as well as your edits, so share it only when you intend to share that working material. The editor does not automatically save projects to your account or browser storage. Keep the tab open until your export or project download has finished. Image processing and OCR run locally; the site's existing analytics and advertising scripts still load separately.
বাংলায় দ্রুত ব্যবহারবিধি
- Choose image দিয়ে JPG, PNG বা WebP ছবি খুলুন। প্রথমে পরিষ্কার, সোজা ও যথেষ্ট বড় অক্ষর আছে কি না দেখুন।
- ছবিতে বাংলা থাকলে বাংলা বা বাংলা + English বেছে Detect text চাপুন। শনাক্ত হওয়া শব্দ বা লাইনের box-এ click করুন।
- নতুন লেখা লিখে font, size, color ও background repair ঠিক করুন। প্রয়োজন হলে নিজের matching font upload করুন। ছবিতে font-এর আসল তথ্য থাকে না, তাই মিল আনুমানিক হতে পারে।
- সাদা বা একরঙা background-এ Sample surrounding color ব্যবহার করুন। হালকা gradient-এ Edge blend দেখুন। জটিল photo-তে clone brush দিয়ে পরিষ্কার অংশ থেকে pixels কপি করুন।
- OCR ভুল করলে Select text area দিয়ে নিজে এলাকা নির্বাচন করুন। শুধু পুরোনো লেখা সরাতে replacement ফাঁকা রেখে repair apply করুন।
- কাজ শেষে Export image দিয়ে ছবি নিন। পরে layer edit করতে চাইলে আলাদা করে Save project করুন। Crop/resize/rotate করলে layer pixels-এ মিশে যায়; ভুল হলে Undo করুন।
বাংলা যুক্তাক্ষর, নাম, সংখ্যা ও তারিখ নিজে যাচাই করুন। পরিষ্কার background-এ ভালো ফল পাওয়া সহজ; ঝাপসা লেখা বা জটিল texture সম্পূর্ণ নিখুঁতভাবে মেরামতের নিশ্চয়তা নেই।
Troubleshooting and frequently asked questions
Can I click any existing word and edit it immediately?
Run Detect text first so the editor knows where the words are. Then click a detected box. If detection misses a word, use Select text area. A flat image has no original text objects to click before recognition or manual selection.
Why does the replacement font look different?
OCR reads letters and their approximate positions; it does not recover the original font file. Choose a closer family, upload a matching font, adjust weight and size, and compare at 100%. Exact matching is not guaranteed.
Why can I still see traces of the old letters?
The selected area may be too small, the cleanup padding may be insufficient, or the sampled background may contain another character. Adjust the region and padding, change the repair method, or clean the area using the clone brush.
Why did the picture behind the text become smooth?
Solid-color and edge-blend repair estimate a background; they do not know the texture hidden by the lettering. Undo the change and use a better clone source for a detailed photograph.
Does a saved image keep editable or selectable text?
No. Image export contains raster pixels. Save a project if you want to edit the editor's layers later. OCR can extract text again, but it does not make PNG or JPG text selectable like a document.
Can this tool edit a PDF directly?
This page edits image files. The separate PDF Editor adds annotations and organizes PDF pages; it is not a universal original-body-text editor. Use the tool whose output format matches your task.
Does the editor upload my image?
The editor code processes images and OCR in your browser and does not send them to an editing API. Language models and script libraries are downloaded from CDNs. The site's existing Google Analytics and advertising scripts are retained.
