Volver a actualizaciones
Nuevo releaseAug 15, 2026

AperiSolve v3.7.7

Plataforma automatizada de estegoanálisis que detecta datos ocultos en imágenes mediante 16 analizadores paralelos, visualización de capas de bits y una wiki integrada para investigaciones CTF y forenses.

Compartir

Aperi'Solve

Website CI Lint Docker Python Version Ruff uv ty License GitHub Issues GitHub Pull Requests Last Commit Contributors Rawsec's CyberSecurity Inventory

Aperi'Solve screenshot

Try it now: https://www.aperisolve.com

Support Aperi'Solve:

Sponsor Aperi'Solve on GitHub

Aperi'Solve is an open-source steganalysis web platform that performs automated analysis on images to detect and extract hidden data using common steganography tools and techniques.

Key features

  • Visualize each bit layer (LSB and other layers) per image channel (R/G/B/Alpha).
  • Color remapping (random palette remaps with 8 generated variants)
  • Runs 16 analyzers in parallel and displays their output, with per-tool success/no-result badges and one-click download of extracted files:
  • In-app wiki (/wiki): a HackTricks-style steganography handbook — a triage methodology, a decision-tree cheatsheet, technique pages per medium (image, audio, text, files), and a guide per analyzer, all authored as Markdown.
  • Internationalization: UI available in English, French, Spanish, German, Russian, Chinese and Portuguese under language-prefixed URLs (/fr/, /es/, …).
  • "Ask AI" panel on the result page: send the analysis results and images straight from your browser to your own LLM provider (bring your own key) for a verdict on what is hidden and what to try next. Two checkboxes pick what the model gets: the images (on by default) and a text matrix of every image's pixel values (off by default; native resolution, so text-only models can read LSB data too). The matrix comes from the browser's image decoder: it is exact for opaque pixels of a PNG/BMP/GIF, but a semi-transparent pixel can be off by one in a colour channel (the canvas stores premultiplied alpha), the RGB under a fully transparent pixel reads 0, and JPEG depends on the decoder.
  • MCP server at https://www.aperisolve.com/mcp (docs/mcp.md): lets Claude, Cursor or any MCP client upload a file and read its analysis.
  • Worker queue architecture for offloading heavy/slow analyzers (Redis + background workers).
  • Content-addressed result cache, HTTP caching, and per-endpoint rate limiting.
  • Results stored temporarily for later browsing and download, then cleaned up automatically by a scheduled job.

Adding an analyzer

Adding an analyzer is a single file: subclass SubprocessAnalyzer, set a few class attributes (name, has_archive, needs_password, deep_only, display_order), and install the tool in the Dockerfile. The worker run list, download allow-list, and frontend order are all derived automatically. See docs/adding-analyzer.md.

Quick start (Docker)

In case you want to host your own version of https://www.aperisolve.com/.

Required: Docker + Docker Compose.

git clone https://github.com/Zeecka/AperiSolve.git
cd AperiSolve
cp .env.example .env
docker compose up -d

Then browse url: http://localhost:5000/

Ads and analytics are disabled by default; every integration (CUSTOM_EXTERNAL_SCRIPT, GOOGLE_ADS_TXT, SITE_BASE_URL, Sentry) is opt-in via environment variables documented in .env.example. The app never injects ad markup itself: CUSTOM_EXTERNAL_SCRIPT loads the AdSense script and placement is left to Auto ads.

Architecture

The stack runs as Docker Compose services:

  • web — Flask app served by gunicorn (Jinja templates + vanilla JS frontend).
  • worker — an RQ worker that runs the analyzers for each submission, fanning out to one thread per tool.
  • cron — an RQ cron scheduler that runs the retention cleanup off the request path.
  • initdb — a one-shot service that creates the tables and pre-fills the IHDR CRC recovery database.
  • postgres — stores image metadata and submission status.
  • redis — RQ broker (DB 0) and rate-limiter storage (DB 1).
  • rqdashboard — queue monitoring, bound to localhost:9181.

This separation keeps heavy tools (binwalk, foremost, zsteg, etc.) isolated and avoids blocking the web worker. Identical submissions are deduplicated by a content hash, and derived images are cached with long-lived immutable HTTP headers, so repeat traffic is cheap.

Documentation

Testing

uv sync --extra dev
uv run pytest -q            # unit tests (no services needed)

# end-to-end analyzer tests through the running app:
docker compose -f compose.dev.yml up -d
uv run pytest tests/test_webapp_analyzers.py -v

Roadmap

See Issues.

Contributing

See CONTRIBUTING.md.

Security

Categorías