Perspective Fix — Free Online Perspective Corrector & Document Enhancer | Run Local Tools

Straighten angled document photos with editable corner handles, best-effort edge suggestions, A4 and Letter presets, scan-enhancement filters, and local PNG, JPEG or WebP export. Free and private.

What it does

Perspective Fix turns a selected four-corner area into a straight rectangular image. Use it for photographed pages, receipts, posters, whiteboards, labels, screens and artwork that look skewed because the camera was off to one side.

Drag four corner points or move them precisely with the arrow keys; Shift + arrow moves a point farther.
Try the best-effort automatic edge suggestion, then inspect and adjust every corner manually.
Choose automatic output dimensions or A4, US Letter, square, and custom pixel dimensions.
Rotate an image by 90 degrees before correcting it.
Improve readability with Original, Enhance paper & text, Grayscale, Black & white, and Eco print presets, plus brightness and contrast adjustments.
Choose a white, black, transparent or custom-colour background.
Export PNG, JPEG, WebP, or a single-page PDF with adjustable image quality.
Undo and redo corner changes, zoom for placement, and compare the original image.

How it works

The tool maps the selected quadrilateral to a rectangle using a projective transformation and interpolates pixels in your browser. The edge suggestion uses a lightweight local edge-projection heuristic; it works best when the page contrasts with its background and is not a guaranteed computer-vision detector. Always check the suggested corners. Export is limited to 16 megapixels and the shared browser canvas dimension ceiling to help avoid excessive memory use.

How to use it

Choose an image, drop it onto the preview, or paste an image from your clipboard.
Optionally rotate the image, then select Auto-detect edges. If the suggestion is wrong or unavailable, drag the four points onto the object’s corners.
Select an output proportion preset or enter custom dimensions.
Choose a scan-enhancement preset and adjust brightness or contrast if needed.
Select the background, output format, quality and filename.
Select Correct & export, inspect the result, then download it.
Use Clear saved work on a shared device to remove the stored source image and settings.

FAQ

Does the image leave my computer?

No. Image decoding, edge suggestions, corner adjustment, filters and perspective correction run locally in your browser. The tool does not upload images or use external image-processing services.

Is automatic edge detection always accurate?

No. It is a best-effort helper based on edge projections, not a trained document-detection model. It can miss low-contrast, curved, cluttered or strongly rotated pages. Check the four points and adjust them manually.

Which image formats can I use?

Use formats your browser can decode, commonly JPEG, PNG and WebP. Export options are PNG, JPEG and WebP.

Can I set a paper size or exact output dimensions?

Yes. Choose A4, US Letter, square, automatic dimensions, or enter custom pixel dimensions. Paper presets are proportional pixel presets, not a print-DPI guarantee.

Does it support PDF or batch processing?

It can export the corrected image as a single-page PDF. Batch queues and multipage PDF export are not included in this version; correct and export one image at a time.

Are my edits restored after refreshing?

The source image and corner/settings state are saved locally in this browser and restored when the tool opens again. Browser storage can be cleared or evicted, so keep an original copy of important files.

Privacy

This tool is local-first. Images are decoded and transformed in your browser; they are not uploaded to a server, and no optional internet feature is included. The source image and working settings are persisted in this browser so work can survive a refresh. Clear saved work to remove them. Browser storage can be cleared by the browser or evicted under storage pressure, so keep a copy of important originals.

Supported formats

Input: browser-decodable image formats; JPEG, PNG and WebP are recommended.
Output: PNG, JPEG, WebP and single-page PDF.
Transparency: PNG and WebP support transparent backgrounds. JPEG and PDF require a solid background.

Limitations

Requires a modern browser with Canvas, ImageData and IndexedDB support.
Automatic edge suggestions are heuristic and should always be checked manually.
No batch queue, multipage PDF export, OCR, curved-page flattening, or trained ML edge detector in this version.
Output dimensions are limited to 16 megapixels and the shared browser canvas dimension ceiling.
Very large source images may use substantial memory and take longer to process.