[BETA TESTERS] Vector Spinner API: A data-driven spinner

I've been working on a completely data-driven spinner, accessible as an API, and I'm currently looking for testers.

How is this spinner different from existing ones? It doesn't use a thesaurus. Instead, similarities in word meaning, rules of grammar, and probabilities of words occurring in a given context are all estimated using state-of-the-art NLP approaches paired with (gigabytes of) training data, currently coming from a variety of free sources, including Wikipedia, product reviews, and PLR articles. This has its ups and downs... but by far the biggest advantage is scalability: improving the average quality is often as simple as throwing more data in.

This means that the quality of results depends on expanding and balancing out the training data - and that's where you come in. I need some field testers to evaluate how well the spinner handles different kinds of content so I can balance out the data it's trained on, as well as get general feedback on quality, usability, and errors.

Who I'm looking for

• members who regularly use spin, preferably if you use existing commercial spinners
• members who have experience with API endpoints/sending out automated requests

What I want you to do

• test out the current (very basic) spinner API (I'll PM/e-mail you the details with the instructions for accessing the API)

Feedback I'm looking for

• reports on any bugs, crashes, and errors
• thoughts on performance: do you find it tolerable, or would you prioritize improving it over other things
• missing vocabulary & niches: notify me of any topics where the vocabulary either isn't spun, or gets replaced with unrelated words (e.g. if vocabulary related to tattoos in your content gets replaced with nonsense or unrelated stuff)
• also: point out any systematic nonsense which gets thrown in (e.g. references to petroleum or Spiderman in articles about fitness)
• general feedback on how the spinner compares with what you're currently using, how well it fares against Copyscape for your type of content, thoughts on the current settings, feature requests, etc.

Interactive demo (with sample articles, just to show off general quality & current customization):
http://jespersen.pythonanywhere.com/vectordemo

Contact (for actually testing the API):
E-mail: [email protected]
Or just leave a reply here or send me a PM if you're interested.

I'm looking for around 10 people, first come first served as long as you think you qualify and you're willing to give serious feedback and respond to updates.

Thread Approved.

Hi,

I do use The Best Spinner regularly for spinning content. so I could test this out for you. i'm not so sure about API and endpoints though. so if you think I qualify, send me a PM.. But i've basic understanding of Python programming so I can learn.

Please consider me for beta testing spot. This really looks like something different and amazing.

If you know Python, that's cool too, since I can give you an example of what to do. Sent you a PM.

Thanks got the PM,

I'll test it out and let you know right away. And also try messing with the python code. This will be nice to play with.

Regards

Hi Jespersen,

I'm interested in your spinner tool. I've used almost all the industry tools related to spinning (from WordAI to thebestspinner). If you think I qualify for the beta testing please PM me.

Thanks

i am interested in testing your spinner tool

Had a brief play with the demo, I like where you're heading with this - I use The Best Spinner a little, but found your tool very user friendly. Not sure I'd be any good to you as a tester, but, if you're ever putting together a list for pre-launch awesome offers, please add me! Best to you for this project.

I'm still looking for more testers if anyone's interested.

I updated the spinner to rewrite expressions with nouns at phrase level (using a technique inspired by machine translation), making it faster, simpler, and more readable, so if you weren't convinced before, check out the updated demo. I want to fix minor details and balance out training data before I focus on extending this to all parts of speech (which should help both readability and uniqueness a lot), so if there's anyone still interested, let me know if you want to give the API a try.

I would like to test this. Please send me the details.

i have been using WAC for a long time now , i would like to be a beta tester please count me in

I sent you a PM. As I said I know some of NLP and coding some scripts won't be an issue for me. I could probably do some fuzzing-oriented tests to your tool and search for bugs/flaws.