PNG to JSON Converter: Free, Private, No Upload

Convert PNG files to JSON in your browser. Smaller output file, batch up to 20 files at once, no upload, no account, no watermark. Your files never leave your device.

PNG to JSON
Batch up to 20 files
No upload, no account
Metadata stripped

1Add up to 20 images
2Pick from 18 formats
3Download or copy the code

Processed on your device. Your images are never uploaded, and metadata is stripped. How we keep it private

How do I convert PNG to JSON for free?

Drop your PNG files into the tool above, choose JSON as output, and download the result. Everything runs in your browser, so nothing is uploaded and there is no limit or watermark. It works on Windows, Mac, Linux, ChromeOS, iOS and Android — anywhere you have a modern browser.

Price: Free, no credits and no daily cap
Account required: No
File uploaded to server: No, processed locally in your browser
Batch limit: Up to 20 files at once
Input format: PNG
Output format: JSON
Best for: JSON workflows, smaller files, faster sharing
Works on: Chrome, Edge, Firefox, Safari, iOS and Android browsers
Privacy: No upload, no tracking, metadata stripped
Output quality: Adjustable 1-100, defaults to 90

How to convert PNG to JSON

1
Drop your PNG files
Drag up to 20 PNG files onto the page or click to choose them. Everything runs in your browser.
2
JSON is preselected
Output is already set to JSON. Adjust quality, resize or batch if you want.
3
Download
Save each JSON, or download all of them together as a ZIP.

PNG to JSON: what changes and when it makes sense

PNG is a Portable Network Graphics format. It is best known for exact pixel reproduction, crisp text and edges, and full alpha transparency. The main drawbacks are large files for photographs and no lossy mode for trading quality against size.

JSON is not an image format. It is a JSON value that wraps the image data as text. Its main appeal is carrying an image alongside other fields in a single payload. The main drawbacks are a size about a third larger than the original and slow parsing for large payloads.

PNG vs JSON at a glance

Property	PNG	JSON
Type	Raster image	Text encoding of an image
Compression	Lossless (DEFLATE), with an optional 8-bit palette mode	The image bytes are written out as Base64 text, which adds about 33 percent to the size
Transparency	Full alpha channel with soft edges	Same as the embedded image
Animation	No	Same as the embedded image
Colour depth	8 or 16 bits per channel, or 8-bit indexed colour	Same as the embedded image
Support	Opens in every browser and image editor	Readable by every language and API client
History	Created in 1995 as a free, patent-unencumbered replacement for GIF after Unisys began charging royalties for LZW compression. Became a W3C Recommendation in 2003. The signature bytes are 89 50 4E 47 0D 0A 1A 0A — "PNG" with newlines designed to catch file corruption.	JSON (JavaScript Object Notation) was specified by Douglas Crockford in 2001. Embedding images as base64 strings in JSON became common with REST APIs and AI model inputs (OpenAI vision, Anthropic Claude, Google Gemini).

What actually changes in this conversion

JSON is an encoding of the image data, not a new compression scheme. The text is about a third larger than the binary file, so what you gain is portability rather than a smaller size.
Transparency follows the image data being encoded. JSON itself is only a text wrapper.
EXIF, GPS and camera data are stripped on every conversion. That is good for privacy, but you lose capture dates.

About PNG: history and technical details

Created in 1995 as a free, patent-unencumbered replacement for GIF after Unisys began charging royalties for LZW compression. Became a W3C Recommendation in 2003. The signature bytes are 89 50 4E 47 0D 0A 1A 0A — "PNG" with newlines designed to catch file corruption.

Technical: Uses DEFLATE (LZ77 + Huffman) compression. Two main modes: 8/16-bit truecolour with alpha, and 8-bit indexed with up to 256 colours. Filters per scanline (None, Sub, Up, Average, Paeth) before DEFLATE for better ratios.

Real-world size: A 1920x1080 screenshot is 200-600 KB as PNG. As JPG at quality 85 it is 100-300 KB but with blurry text. A 12-MP photo is 18-25 MB as PNG, 2-4 MB as JPG. PNG wins on text, JPG wins on photos.

Popular tools: Adobe Photoshop, GIMP, Figma, Sketch, Snipping Tool, macOS Screenshot, ImageMagick, pngquant, oxipng.

Alternatives: WebP lossless (20-30% smaller), AVIF (better for photos), JPG (photos only), SVG (vector graphics).

About JSON: what you get

JSON (JavaScript Object Notation) was specified by Douglas Crockford in 2001. Embedding images as base64 strings in JSON became common with REST APIs and AI model inputs (OpenAI vision, Anthropic Claude, Google Gemini).

Technical: { "image": "iVBORw0KGgo...", "name": "photo.png" }. The base64 string can be escaped differently per language. Some APIs require no prefix, others require data:image/png;base64, prefix.

Real-world size: A 1 MB image as JSON base64 is ~1.33 MB. Parsing 10 MB JSON is 100-500ms in most languages. This is why vision AI APIs often recommend URL uploads for large images.

Popular tools: OpenAI API, Anthropic API, Google Gemini API, Postman, any JSON parser.

Alternatives: Multipart upload, Object storage + URL, Signed URL, Base64 in headers (rare).

Browser and software compatibility

PNG: Opens in every browser and image editor. JSON: Readable by every language and API client.

JSON with an encoded image works with any parser. Large payloads parse slowly.

Why people move away from PNG

A PNG of a photograph is commonly several times larger than the same image as JPG or WebP, which hurts page weight.
Screenshots and photos exported as PNG are among the heaviest assets on many sites.
Some upload forms and print services reject PNG or downscale it aggressively.

Why people choose JSON

JSON with an encoded image is useful for API fixtures, config files and mock data.
The image travels together with the other fields as a single payload.

Common mistakes with PNG

Using PNG for photographs — files are 5-10x larger than JPG with no visible benefit
Saving screenshots as JPG by mistake — text becomes blurry and unreadable
Forgetting PNG transparency — white backgrounds appear where there should be alpha

Best practices for keeping quality

PNG sources often contain transparency. Decide before converting what should replace it: white, black or a brand colour.
Large payloads are slow to parse and send. Store files in object storage and keep a URL in the JSON for anything over a few hundred KB.
Resize before converting if the image will be shown small. A 4000 pixel photo displayed at 800 pixels wastes bytes in any format.
Check the result at 100 percent zoom, especially around text, edges and gradients, before you replace any originals.
Keep your original files. Conversions to a lossy format cannot be undone.
When you convert a set of images, use the same quality setting for the whole batch. The converter handles up to 20 files at once.

PNG to JSON questions

Quick answers about privacy, quality, supported formats, file limits and how JSON works.

Why convert PNG to JSON?

JSON files are typically smaller than PNG, which makes them faster to share, send and load. The trade-off is a small amount of image detail at low quality settings.

Will converting PNG to JSON lose quality?

At 90% quality, most people cannot tell the difference. Use 85-95% for photos and you are fine.

Does JSON keep transparency?

JSON is only a text wrapper around the image data, so transparency depends on the image format inside it.

Are my files uploaded?

Never. Everything runs in your browser. Your files never leave your device. Open DevTools Network tab while processing — you will see no upload request.

Can I convert many files at once?

Yes, up to 20 files at once. Download them individually or all together as a ZIP.

Does the converter strip metadata?

Yes. Browsers do not pass EXIF data through a canvas, so GPS, camera model and dates are removed on every conversion. That is good for privacy.

Is it free?

Yes. No credits, no daily cap, no account, no watermark. The practical limit is your device memory.

Do AI vision APIs accept base64 images?

Yes. OpenAI, Anthropic, and Google all accept base64 images in their API requests, either as a raw string or as a data URL depending on the provider. Check the specific API docs for the exact format.

Your files never leave your device

No upload, no account, no tracking. Everything happens in your browser using the Canvas API and WebAssembly. You can confirm this in your browser's DevTools Network tab while the tool runs — no request carries your file. See why ToolDeck is different.