Hide any file inside an image — entirely in your browser.
No uploads. No server. No build step. Just open index.html.
StegoLab is a client-side steganography tool: it hides an arbitrary file (a document, another image, a ZIP, audio, plain text — anything) inside a carrier image, so the output looks like an ordinary picture but secretly carries your payload. Everything happens locally in the browser using the Canvas API and the Web Crypto API — nothing is ever sent to a server.
Under the hood it's more than basic LSB embedding: payload bytes can be GZIP-compressed and AES-256-GCM encrypted before Sobel edge detection picks the highest-detail pixels to hide them in, and a password-seeded PRNG scrambles the exact embedding order so the layout can't be reconstructed without the password.
- Hide any file type — text, PDF, ZIP, images, audio, or anything else, inside a PNG, BMP, TIFF, or WebP carrier image
- Optional AES-256-GCM encryption — password-based, with PBKDF2 key derivation (200,000 iterations) and a random salt/IV per file
- Automatic GZIP compression — applied before encryption when it actually reduces size
- Adaptive LSB embedding — Sobel edge detection ranks pixels so data is hidden in high-detail regions first, where changes are least visible
- Password-seeded PRNG — randomizes embedding order and channel rotation (R/G/B) so the layout can't be recovered without the password
- Capacity & security stats — live capacity estimate while picking a payload, plus modified-pixel count and PSNR after encoding
- Hex viewer — inspect a decoded payload's raw bytes
- Dark/light theme, drag-and-drop or click-to-browse for both the carrier image and the payload file, keyboard shortcuts
- Open
index.htmlin a modern browser (Chrome, Firefox, Edge, or Safari). - Encode: pick a carrier image and a payload file (or type a message), optionally set a password, then click Encode. A PNG with the hidden data downloads automatically.
- Decode: pick the encoded image, enter the password if it was encrypted, then click Decode to recover the original file.
StegoLab/
├── index.html
├── styles.css
├── LICENSE
├── app/
│ └── StegoLab Setup 1.0.0.exe
└── js/
├── app.js # wires up the UI and orchestrates encode/decode
├── core/
│ ├── constants.js # magic bytes, flags, header layout, crypto params
│ ├── header.js # binary header build/parse
│ ├── encoder.js # full encode pipeline
│ └── decoder.js # full decode pipeline
├── crypto/
│ └── aes.js # AES-256-GCM encrypt/decrypt (Web Crypto API)
├── compression/
│ └── gzip.js # GZIP via CompressionStream
├── steganography/
│ ├── prng.js # Mulberry32 PRNG, password-seeded
│ ├── sobel.js # edge-magnitude computation
│ ├── embedder.js # adaptive LSB embedding
│ └── extractor.js # LSB extraction (mirrors embedder)
├── ui/
│ ├── dashboard.js # tab switching
│ ├── theme.js # dark/light theme + persistence
│ ├── animation.js # full-screen encode/decode animation overlay
│ ├── progress.js # progress bar during encode/decode
│ ├── dragdrop.js # drag-and-drop + click-to-browse for images
│ ├── keyboard.js # keyboard shortcut registry
│ └── stats.js # image stats, capacity, security analysis
├── utils/
│ ├── crc32.js # payload checksum
│ ├── helpers.js # byte/format helpers
│ ├── bundle.js # payload bundling helpers
│ └── file.js # file reading, image decoding, MIME/type detection
└── storage/
└── settings.js # small localStorage key-value wrapper
┌─────────────────┐
│ Carrier Image │
│ (PNG/BMP/TIFF) │
└────────┬────────┘
│
┌────────────┐ ┌───────────┐ ┌──▼──────────┐ ┌────────────┐ ┌──────────┐
│ Payload │────>│ GZIP │────>│ AES-256- │────>│ Sobel │────>│ Adaptive│
│ (any file)│ │ Compress │ │ GCM Encrypt│ │ Edge │ │ LSB │
└────────────┘ └───────────┘ └─────────────┘ │ Detection │ │ Embed │
└────────────┘ └────┬─────┘
┌──────────┐ ┌────────────┐ ┌────────────┐ │
│ Binary │<────│ Header │<────│ PRNG │<─────────┘
│ Output │ │ (magic, │ │ (password │
│ PNG │ │ flags, │ │ seeded) │
└──────────┘ │ CRC32...) │ └────────────┘
└────────────┘
┌──────────┐ ┌────────────┐ ┌────────────┐
│ Encoded │────>│ Header │────>│ PRNG │
│ Image │ │ Parse │ │ (password │
└──────────┘ └────────────┘ │ seeded) │
└──────┬─────┘
┌────────────┐ ┌───────────┐ ┌─────────────┐ │
│ Original │<────│ GZIP │<────│ AES-256- │<─────────┘
│ Payload │ │ Decompress│ │ GCM Decrypt│
└────────────┘ └───────────┘ └─────────────┘
Header: a small binary header (magic bytes, version, flags, filename, MIME type, payload length, timestamp, CRC32, and — if encrypted — salt and IV) is embedded sequentially in the image so it can always be parsed, even without a password.
Payload placement: Sobel edge detection scores every pixel, and the highest-detail pixels are used first. Within that set, a PRNG seeded from the password (or a fixed seed if none is given) shuffles both the pixel order and the R/G/B channel used for each bit, so the embedding pattern can't be reconstructed without the password.
Encryption/compression: payload bytes are GZIP-compressed (skipped if that doesn't help), then optionally encrypted with AES-256-GCM using a PBKDF2-derived key, before being embedded.
Requires Canvas, the Web Crypto API, and (for compression) CompressionStream — supported in current Chrome, Firefox, Edge, and Safari. If CompressionStream isn't available, compression is silently skipped.
MIT — see LICENSE.