cracker0dks/CaptchaSolver
JDownload 2 Offline Captcha Solver
What it is
CaptchaSolver is a small add‑on for JDownloader 2 that lets the download manager automatically solve a handful of image‑based captchas that it otherwise asks the user to type in. It does this by bundling a pre‑trained YOLO object‑detection network (via the Darknet framework) and a thin Node.js wrapper that feeds the captcha image to the network and reads the predicted characters.
Who might use it
- Power users of JDownloader who frequently download from the listed file‑hosting sites and want to avoid the manual “type the letters” step.
- Self‑hosted JDownloader instances (NAS, home server, Linux box, macOS) where a fully‑automated download pipeline is desirable.
- Anyone comfortable installing Node.js, compiling Darknet, and tweaking a few shell scripts.
How it works (high‑level)
- Capture – JDownloader saves the captcha image to a temporary file.
- Invoke Darknet – The Node.js script calls the compiled Darknet binary, loading a YOLO model that has been trained on the specific captcha styles used by the supported hosts.
- Predict – Darknet returns the most likely characters with confidence scores.
- Return to JDownloader – The script formats the prediction and feeds it back to JDownloader, which proceeds with the download.
The repository ships the necessary Darknet binaries for amd64 (x86_64) and provides instructions for compiling it on other architectures.
Installation steps (summarized)
- Windows – Download the provided zip, copy the
JDownloader 2.0folder into your existing JDownloader installation, and restart. - Linux – Install Node.js, clone the repo, compile AlexeyAB’s Darknet (or use the pre‑built binary), copy the
darknetexecutable intotools/offlineCaptchaSolver/darknet64/, then copy theJDownloader 2.0folder into~/.jd(or the Flatpak path) and restart. - macOS – Install Node, CMake, OpenCV via Homebrew, build Darknet with OpenCV support, then copy the repo’s files into the JDownloader app bundle.
The README also explains how to deactivate individual host‑specific solvers if a site changes its captcha format.
What it can solve today
| Host | Captcha type |
|---|---|
| keep2share.cc / k2s.cc | 6‑digit alphanumeric |
| fileboom.me / fboom.me | 6‑digit |
| tezfiles.com | 6‑digit |
| publish2.me | 6‑digit |
| filejoker.net | Geometrical (shapes) |
| depositfiles.com / dfiles.eu | 6‑digit |
Older captcha formats (4‑digit, legacy 6‑digit) are no longer supported in the current version but are still available in the v1.x release.
Limitations & maintenance
- Scope limited to the hosts listed – adding a new site requires training a new YOLO model and adding a wrapper script.
- CPU‑only inference – the bundled Darknet binaries run on the CPU; performance depends on the host’s processor.
- Manual setup – users must compile Darknet for non‑amd64 platforms and adjust script paths (e.g.,
node/npmlocations) on macOS. - No automatic updates – the project is a niche utility; updates are released as new JDownloader versions or when a host changes its captcha.
When you might look elsewhere
If you need a general‑purpose captcha‑solving service that works across many sites, supports modern reCAPTCHA v2/v3, or offers cloud‑based APIs, a dedicated service (e.g., 2Captcha, DeathByCaptcha) would be more appropriate.
All details above are taken directly from the repository’s README.
Related
- Project
- Project
- Project
- Project
- Project