Our API documentation states that these notifications are optional. Captcha recognition will still work without it. But if you consider sending these notification captcha recognition will be more precise and even cheaper.
Reportbad used with GET/POST request like
This request should be sent every time when the site you are solving captcha for didn't accept our resolution and said it's incorrect.
What happen after you send this notification:
- For simple captcha/image captcha
Additional resolution of this captcha is initiated for other worker with high rating. Then we compare resolution results of these two workers (case sensitive if it was requested). If answers are equal we decide the complain was in fact incorrect and both workers are rewarded for this captcha resolution. If answers are NOT equal customer receives a refund and first worker incorrect resolutions count increased by 1. Then we additionally check his answer with moderator's resolution and if it differs too much we re-check several more of this worker's answers. Depending of the results this worker could be banned. You could come to a conclusion that we use our resources ineffectively (including finances, as every answer counts and sometimes one captcha costs more than we've been payed by customer). But in a long run with big amount of captchas this approach is justified.
- For captchas with token: ReCaptcha V2, KeyCaptcha, FunCaptcha
There is no way to re-check this type of captcha. However, we gather reportbad statistic for every worker and if a worker receives too many reportbad messages or no reportgood notifications, we limit his work with this type of captchas. We additionally recognize this worker linked accounts and limit these accounts too.
- ReCaptcha V3
Due to specific of ReCaptcha V3 answers of one workers could be good for one site and bad for another. When we receive reportbad for a worker we lower his rating for a particular domain. In case of low rating for all of domains we limit his work with ReCaptcha V3.
Reportgood used with GET/POST requests like
This request should be sent every time if you are sure that the resolution provided was correct.
Why is it so important to send reportgood notifications:
- For normal captcha/image captcha
We have a rule that we don't block the worker for incorrect answer if he was trying to answer correctly. He will also get paid for it. If a worker makes some tipical mistakes he will be sent back to training by moderators (and yes we do have training and examination for workers). However if we see that a worker makes deliberate mistakes we will block him. Most of the workers are working hard and diligently to solve all kind of captcha correctly even when it's hard. But this policy makes us no good as now there are some workers who doesn't care if it's O or 0, small l or capital I. And those workers receive same payment as good workers which is not fair. So if we have enough statistic regarding good answers we can calculate worker earn coefficient and pay more to those who work hard and diligently.
- For captchas with token: ReCaptcha V2\V3, KeyCaptcha, FunCaptcha
Workers who provide good answers will be prioritized over others to receive these types of tasks and earn more (higher coefficient system is not implemented in current model for these types of captchas).
If you start sending reportbad\reportgood notifications:
- You — will receive refunds for bad answers and overall quality of answers will be higher
- We — will detect bad workers faster
- Workers — will earn more!