[feature] Implement AntiCaptcha or FlareSolverr support

[JReming85]

Is your feature request related to a problem? Please describe.
There are a few sites that ident my browser as a bot, I have tried using residential ip's but it will still get flagged most of the time.

Describe the solution you'd like
I currently run my own Captcha solver, XEvil 6.0 which can handle pretty much all types of captchas and currently set to emulate AntiCaptcha's API.

If possible would love to see it get implemented, Implement AntiCaptcha's Plugin w/ Selenium

I even modified their plugin to use my server, would even be willing to provide an api key and plugin to help the devs with implementing. XEvil supports AntiCaptcha, 2Captcha, De-Captcher, iMarcos, RipCaptcha, RuCaptcha, and DeathByCaptcha so others can also be implemented, not to mention quite a few of the others will either use the same API setup as AntiCaptcha or 2Captcha.

There is also another project I use, mostly for Jackett to auto search torrent sites called FlareSolverr. It will attempt to bypass the hCaptcha's from Cloudflare (Captcha plugin solves them as well, but not currently supported by FlareSolverr) and output the site in a json response.

Would also love to see options to auto save cookies, for a more persistent experience. One of the current ways I am getting around the captchas is by solving it via a private browser and copy/pasting the cookies into changedetection however at some point I will most likely have to replace/update them manually.

Describe the use-case and give concrete real-world examples
Think I outlined it pretty well in the previous section but would like to see

1. Captcha solver plugin implemented when a captcha is detected
2. Ability for it to self update/store cookies

Additional context
If you wish to try my captcha plugin or need a api key to test, let me know (devs only).

[dgtlmoon]

That's pretty interesting, I never knew such technologies exist

One thing tho is that I see selenium as essentially deprecated when compared to Playwright+Browserless, Browserless has significantly more work added to it around evading bot detecting, and is just in general a better system to use (for example, selenium cant even take full page screenshots)

it seems that anti-captcha is a paid service tho right? https://anti-captcha.com/clients/entrance/login

[JReming85]

Yes, most captcha solvers are paid but there are software you can purchase to do the work for you.

The current best being XEvil 6.0 (beta), followed by CapMonster. I currently have a license for XEvil 6 that can potentially solve all text, math, google, and hcaptcha. More then happy to provide access so there will be no out of pocket captcha solving costs on the dev teams part and I can adjust XEvils API to emulate the other paid services.

There are also other captcha solutions that emulate the larger services but are much cheaper like https://capmonster.cloud/en/ which can support most of the ones my software can.

Eitehr way if your interested, more then happy to send you q api key to my server http://captcha.jremi.com or https://captcha.jremi.com

Also if a plugin is not the best move, my friend created automated software to claim crypto faucets https://faucetcollector.com/

They way he handled it, was the bot stored the endpoint/url of the server + the api key. From there each site was scripted in C# and had a section dealing with captchas.
https://faucetcollector.com/docs/index.html#captcha-solving

For best results, you would provide the xpath for the captcha, like to text/math captchas, xpath to the image and text box. For recaptcha xpath to the recaptcha itself.

Might be able to implement something similar, if xpath found then solve captcha or something

[dgtlmoon]

I wonder if things like mechanical-turk are still going and you could send a screenshot of the element and wait for the solution

[JReming85]

Thats an option if you wanted to start your own service, similar to how 9kw.com works.

You solve captchas to earn credits to get your captchas solved by others.

[dgtlmoon]

moved to discussions (atleast for now)

[dgtlmoon]

if its going to be a paid solution, then there are many possibilities like https://brightdata.com/webinar/solve-and-prevent-recaptcha

[JReming85]

Unfortunately as I originally stated, those are not working for a few sites for me most of the time

[dgtlmoon]

I'm open to PRs I guess :)

[JReming85]

and about the 2nd part with the self updating cookies?

e.g. one of the sites that does not work is
https://shop.asus.com/us/90yv0ie1-m0aa00-tuf-rtx4090-24g-gaming.html

I currently have to open in a private browser, solve the captcha and copy the cookies every day, when just cloning the cookie from the last visit "should" solve me having to manually touch it every 24h.

[dgtlmoon]

I dont know the answer, sorry

[semaf]

Brightdata is not solving for my case Google Recaptcha

[2fst4u]

Flaresolverr would absolutely be a plus. I use it for all my private trackers and it works great and is acutely maintained.

[soleous]

I'm not adapt at FlareSolverr, but I've noticed its using Selenium. So does it work out of the box?

[mat926]

There is also https://github.com/dessant/buster which solves Captchas from audio challenge.

[dgtlmoon]

Update FlareSolverr - 2025 looks like this project as dead and cloudflare got too hard for it

[SunRed]

Just as a heads-up, we have Byparr now and in my testing it reliably bypasses Cloudflare challenges.

[dgtlmoon]

@SunRed it would help ALOT if that byparr project bothered to write a simple "how to" on their readme

[maxoux]

I think it's the same api as in flaresolverr, since byparr is a as-is replacement used with prowlarr
Would be very neet to have it compatible tho, have more and more website behind cloudflare

curl -L -X POST 'http://localhost:8191/v1' \ -H 'Content-Type: application/json' \ --data-raw '{ "cmd": "request.get", "url": "http://www.google.com/", "maxTimeout": 60000 }'

edit: Just checked, this command is indeed working with byparr

[seth100]

I have a FlareSolverr container setup and working, but now my question is: what url should I use in ChangeDetection as "Browser connection URL"?

[hucknz]

cloudflarebypassforscraping is an alternative to FlareSolverr that seems to still be working with Cloudflare.