Four New Image Tools: Cropper, Compress-to-KB, OCR, Photo-to-Sketch

The pattern with most free image tools online is the same: drag in a photo, watch it upload to someone else's server, get a watermark, hit a daily limit, or get nudged toward a paywall. None of that is necessary for any of these tasks. Modern browsers can crop, compress, OCR, and stylize images entirely on-device — the only reason it's usually done server-side is to bill for it.

This batch of four tools all share that posture: the image never leaves your browser. Verifiable in DevTools → Network tab during the operation — zero requests sent. Below: what each tool does and the specific gap it fills.

1. Image Cropper — 7 aspect ratios + free-form

What it does. Drop an image, drag the crop frame, get a cropped JPG or PNG. Pick from 7 standard aspect ratios — 1:1 (Instagram square), 16:9 (YouTube / widescreen), 9:16 (Reels / Stories / Shorts), 4:3, 3:2, 2:3, 3:4 — or free-form. 90° rotation handles sideways phone photos.

The gap it fills. Most online croppers either lock you to a single aspect ratio or upload your photo to fetch “previews.” Pyrelo's cropper uses react-easy-crop rendered to canvas locally — instant preview, no uploads, full output quality.

Try it. /tools/image-tools/image-cropper

2. Compress to Target KB — built for Indian govt portals

What it does. Set the size you need (20 / 50 / 100 / 200 / 500 KB / 1 MB) and the tool finds the highest-quality JPG (or WebP) that fits under that limit. The output is guaranteed within 1-2% of the target — never over.

How. Binary search over JPG quality (0.05 → 0.95). 12 iterations land within bytes of the target. Runs entirely via Canvas in your browser.

The gap it fills.Indian government portals enforce strict file-size limits on every photo and signature upload — and they reject the form if you're even a kilobyte over. UPSC: 20-300 KB. IBPS / SBI / RBI: 20-50 KB. Passport Seva: 10-200 KB. Aadhaar UCL: 100 KB. NEET / JEE: 10-200 KB. State e-district portals: 50-100 KB. Most online compressors let you tweak quality and tell you the resulting size after — this one reverses that: you specify the size, we find the highest quality that fits.

Try it. /tools/image-tools/compress-to-size

3. Image to Text (OCR) — English / Hindi / Tamil and 11 more

What it does. Drop a photo or screenshot, get the embedded text out as copyable plain text. 14 languages supported including English, Hindi (Devanagari), Marathi, Tamil, Telugu, Bengali, Kannada, Malayalam, Gujarati, Punjabi, Urdu, plus French / German / Spanish.

How. Powered by tesseract.js — a WebAssembly build of Google's open-source Tesseract OCR engine. The language data is fetched from tessdata.projectnaptha.com on first use (~10 MB per language), then cached. Inference runs on your CPU; no GPU needed.

The gap it fills. Most browser-based OCR tools are English-only. The few that support Indic scripts usually upload your image to a server. This one runs every language — including Devanagari, Tamil, Bengali — locally.

Try it. /tools/image-tools/image-to-text. Useful for: pulling text out of a scanned PDF page, transcribing a printed letter, lifting text from a screenshot you can't copy from, digitizing handwritten notes (best with clear handwriting).

4. Photo to Sketch — Pencil / Charcoal / Ink

What it does. Convert a photo to a sketch in one of three styles: pencil (soft graphite shading), charcoal (high-contrast smudged effect), or ink (stark line-art). Live preview as you switch styles.

How. Pure Canvas filters — no AI model needed. Edge detection plus tone mapping plus a configurable filter stack (grayscale → invert → blur → blend). The whole pipeline is a few hundred lines of plain JavaScript.

The gap it fills. Photo-to-sketch sites are usually loaded with ads, signups, or watermarks. This one is just a Canvas pipeline that runs in your tab.

Try it. /tools/image-tools/photo-to-sketch. Best results on portraits and product photos with a clear subject and even lighting.

What ties them together

Every one of these tools follows the same operating rule: your image never leaves the browser. There is no upload step, no “preview” round-trip, no metering pipeline. Verifiable: open DevTools → Network during any operation and you'll see zero outbound requests for the image itself. The only network activity is the one-time download of the OCR language data, fetched from tessdata.projectnaptha.com and cached locally afterwards.

That's the unifying frame for the whole /tools suite — every browser-doable task should be browser-done. With these four additions, Pyrelo now ships 32 free in-browser tools across PDF, image, sheet, and website utilities. None of them require an account.

Coming up next — likely candidates from the drafted list: a barcode/QR generator, a PDF form-filler, an audio waveform visualizer. If there's a tool you reach for repeatedly that always involves uploading or signing up, message us — that's the kind of gap we like filling.