[Journey] 1 million UVs/month in 12 months using AI generated content. Let's do it!

[RowdySadjit]

Working!

I managed to make the article builder in like 48 hours. I am so pleased with myself lol.

Still, I will wanna refine it for sure. For the moment I will have to go through and edit the output manually as some bits don't make total sense but that already cuts down 90% of the labor of creating the articles. Will look into the more advanced semantic programming discussed in this thread to make things even more hands free in due course.

 

[Razorbl2]

@satre how many sites you run now?
Have any of your sites received manual penalty?

 

[spectrejoe]

For keywords, do you scrape PAA & use those as your article titles or you do something different to get keywords?

 

[RowdySadjit]

Ok, my kw research bot is up using PAA lib but I face captcha now. How should I avoid it?

I guess using a webdriver is better but the PAA lib is very convenient if captcha can be avoided somehow with it.

I read @Alexion say he does not suffer captchas due to using webdriver and clicking the paa results there meaning less requests but I forgot what the other methods are detailed in the thread. I did a search of thread for 'captcha' but didn't find the results.

What do you guys recommend? There was some solver bot using audio on github. I will look into that. Any other suggestions? I have 10 proxies already that I use for social media but they are low quality data centre and trigger captcha often even without making any requests (they are being abused by others I think).

 

[Alexion]

There are a few options.

1) use your own residential IP, use selenium, get all of the PAA in one request. This is done by clicking on the questions and waiting for new ones to pop up.

The logic is simple. Click on all 4 questions, then click again to close them. Reload the webelements with the new questions, click again two times. Repeat this process 4-5 times, it will provide you with a lot of PAAs.

For relevant questions, it would be best if you click in a backward order, from the last one to the first. Otherwise, you will lose relevancy after the 4th question, as you will only go deeper into other sub categories and losing the main one.

2) Use selenium and reCaptcha solver from github which uses voice recognition. Not easy to implement if you are new to python. I posted a link here somewhere.

3) rate limit that PAA library. I know it kind of floods google the last time I tried it.

4) Use serpApi, but that will be expensive in the long run.

 

[spectrejoe]

There are a few options.

1) use your own residential IP, use selenium, get all of the PAA in one request. This is done by clicking on the questions and waiting for new ones to pop up.

The logic is simple. Click on all 4 questions, then click again to close them. Reload the webelements with the new questions, click again two times. Repeat this process 4-5 times, it will provide you with a lot of PAAs.

For relevant questions, it would be best if you click in a backward order, from the last one to the first. Otherwise, you will lose relevancy after the 4th question, as you will only go deeper into other sub categories and losing the main one.

2) Use selenium and reCaptcha solver from github which uses voice recognition. Not easy to implement if you are new to python. I posted a link here somewhere.

3) rate limit that PAA library. I know it kind of floods google the last time I tried it.

4) Use serpApi, but that will be expensive in the long run.

I started using selenium instead of the PAA module.
It grabs all Questions + Answers but even with 15 proxies it's still getting rate limited somehow which is annoying. I think that every time a new answer is loaded it's a request(loading the questions isn't but loading the answers is I think) hmm

 

[Alexion]

I started using selenium instead of the PAA module.
It grabs all Questions + Answers but even with 15 proxies it's still getting rate limited somehow which is annoying. I think that every time a new answer is loaded it's a request(loading the questions isn't but loading the answers is I think) hmm

I don't know in what volume you're doing this.

In my case, for around 500 seed questions per day, I am able to load up to 80 questions( i stop it after that number ) without facing any captcha. Using my own residential 4G network from my phone. It changes every 3-4 days I think.

In volume, yes, captcha shows up. I posted a link to a resolver. I made it work and it works wonders. It only failed once after 4000 captchas.

 

[spectrejoe]

I don't know in what volume you're doing this.

In my case, for around 500 seed questions per day, I am able to load up to 80 questions( i stop it after that number ) without facing any captcha. Using my own residential 4G network from my phone. It changes every 3-4 days I think.

In volume, yes, captcha shows up. I posted a link to a resolver. I made it work and it works wonders. It only failed once after 4000 captchas.

yea I'm trying to see if I can upload 2000-4000 posts per day so trying to be as fast as possible, assumed proxies would prevent any rate limits, guess 15 isn't enough. I'll take a look at the captcha solver thanks!

 

[Alma]

There are a few options.

1) use your own residential IP, use selenium, get all of the PAA in one request. This is done by clicking on the questions and waiting for new ones to pop up.

The logic is simple. Click on all 4 questions, then click again to close them. Reload the webelements with the new questions, click again two times. Repeat this process 4-5 times, it will provide you with a lot of PAAs.

For relevant questions, it would be best if you click in a backward order, from the last one to the first. Otherwise, you will lose relevancy after the 4th question, as you will only go deeper into other sub categories and losing the main one.

2) Use selenium and reCaptcha solver from github which uses voice recognition. Not easy to implement if you are new to python. I posted a link here somewhere.

3) rate limit that PAA library. I know it kind of floods google the last time I tried it.

4) Use serpApi, but that will be expensive in the long run.

yea I'm trying to see if I can upload 2000-4000 posts per day so trying to be as fast as possible, assumed proxies would prevent any rate limits, guess 15 isn't enough. I'll take a look at the captcha solver thanks!

I can recommend using dedicated residential rotating proxies, they work very well for scraping PAAs.

 

[spectrejoe]

I can recommend using dedicated residential rotating proxies, they work very well for scraping PAAs.

Thanks I'll give it a try!

 

[Niffo]

I just run 4-6 threads on my residential IP. 4-6 on my 5g from phone. And regularly check the laptops for captchas and complete them by hand. No delays, just full speed scraping hehe.

 

[spectrejoe]

I just run 4-6 threads on my residential IP. 4-6 on my 5g from phone. And regularly check the laptops for captchas and complete them by hand. No delays, just full speed scraping hehe.

PAA?
I run 1 thread and get captchas after 15 minutes even on my residential IP lmao

 

[Niffo]

PAA?
I run 1 thread and get captchas after 15 minutes even on my residential IP lmao

I have to add, my residential IP is shared with multiple other people working from home. Doing non-seo related stuff, and I'm in a tier-1 country. So that's probably a lot of the reason why I get such good results.

Also I'm not using the PPA library, but a custom coded solution with selenium.

 

[mdevo]

PAA?
I run 1 thread and get captchas after 15 minutes even on my residential IP lmao

I've been running my own script to scrape PAA's using Selenium from my home IP. Never got captcha tho, so that shounds weird for me. Maybe delay between actions and requests is too small?

 

[spectrejoe]

I have to add, my residential IP is shared with multiple other people working from home. Doing non-seo related stuff, and I'm in a tier-1 country. So that's probably a lot of the reason why I get such good results.

Also I'm not using the PPA library, but a custom coded solution with selenium.

DM'd with how I'm automating with selenium you if that's okay.

I've been running my own script to scrape PAA's using Selenium from my home IP. Never got captcha tho, so that shounds weird for me. Maybe delay between actions and requests is too small?

What delays are you using? I'm only using 1 s delay when loading the google page to make sure everything loads & 0.25s delay between each click on the questions (so 0.25s between each new answer showing). I do open up to 35 answers though per keyword.

 

[Niffo]

Delays are not ur solution, behaving like a human is. Sent u a pm with some more details

 

[Niffo]

U could also try this: https://www.blackhatworld.com/seo/free-create-your-own-ipv6-proxies.1316548/
Haven't tried it myself yet, but google does work with a ipv6 only connection.

 

[mdevo]

DM'd with how I'm automating with selenium you if that's okay.

What delays are you using? I'm only using 1 s delay when loading the google page to make sure everything loads & 0.25s delay between each click on the questions (so 0.25s between each new answer showing). I do open up to 35 answers though per keyword.

Only 1s delay between PAA boxes clicks. No additional delays as long as there is thing called "explicit wait" in Selenium.

 

[spectrejoe]

Only 1s delay between PAA boxes clicks. No additional delays as long as there is thing called "explicit wait" in Selenium.

Yea that's 4x more than I use right now, might be the reason why, although I do use 15 proxies so it should make up for it.

 

[Archemike]

https://textsynth.com/playground.html cool new tool in the gpt realm