saradhi

Congrats on your launch. May I check what is the best email contact to say hi to the tech founder?

From time to time, we receive requests from our clients (mostly from the developed countries) to pipeline humans to the service we offer on internet and provide an error free output for a premium price. While the revenues attached to the human-in-loop triple the profits, the uncertainty in satisfying such requests are a concern to us when employing the data validators full-time or even on demand. How do you deal such scenarios?

Regarding OCR detection, "Sometimes it's not that can't, they don't (may not) want to."

Like many say, validation is basically questioning. 1. Is it useful to you, use it at least 5 times a week? 2. Does anyone else need it? 3. Have you tried the top 10 Google result websites? Are they bad? How better is yours? 4. Try to pitch to your target users through forums not ads.

No, I did not stop my day job to start a new side project or company. The only time I quit my job was the very first time, I started the project. Took 2 years to close it with no big success. And I realized I should have done that without quitting my full time. Since then, I built 6 side projects, never tried to monetize. The latest one was started with an intention to make money, doing good since day 2. Concentrate on SEO if you believe the top 10 Google results of your focus keyword are doing bad. Once you get into top 2, thats all!!

I run a document parsing service, focused on structure detection like tables and form-data. We have tested 27 OCR engines till date, and we do it every 6 months. In our latest evaluation hyperscience.com turned out to be the best in HW recognition, expensive too. Azure's cognitive service is next to it.

P.S: I've no affiliation with the mentioned services

If you own a domain, at 1$/month zoho mail is suggested. Never had a issues with it.

building a form-data extraction API - to extract any form data field-value pairs from images and pdfs; actively working with a couple of existing users.

Edwin, I've reported the incident to you on 9th Oct 2020, never received a solution - not even an acknowledgment.

https://imgur.com/a/DA5rIpC

I guess they support only Indian companies but global payments, luckily ours was Indian registered.

Stripe is not a good choice for Indian payments, we had trouble taking payments as the gateway rejects visa cards too. We've contacted the support multiple times but never received a solution. I recommend to try RazorPay.com (YC backed) just for Indian subcontinent.

I've gone through your submissions to other posts, the work is impressive. Are you available for freelancing? Can you update your profile with email ID or drop me an email (id in my profile)

That was the wrong assumption, as we released a library that imitates to be a pro version from the contributors of open sourcem We sorted it out by unpublishing the library. That's nothing to do with their tech in backend. Feel free to make comparisons on the output quality.

I started a SaaS service [1], in Aug 2019, to extract tables from images and scanned PDFs. Launched with a post on hacker news. I added couple of contractors who share the development on-demand. Last month, for the first time, we hit 10K users, no paid advertising, hitting 170K requests per month. Not expecting to cash out any profits now, investing them in the development. Now, I'm dealing B2B contracts and taking customizations for the businesses. The goal is to hit at least 3K in profit by next year end. Also, releasing a light weight version of the current published version as a new SaaS [2] at low cost.

[1]: https://extracttable.com

[2]: https://imagetoexcel.com

Never tried Fastmail. But it seems Zoho email is comparatively cheaper at 12$/year.

Been using Zoho since 2014 for my side gigs - pretty satisfactory.

May I request you to check our product. That "extra intelligence" is what we are differentiated from all other 99% of pdf to excel services.

With some experience in serving a similar need, the problem is not the lack of resources on the internet but the privacy and security. Be ready to respond to user concerns around it. Occasionally, you may have to handle free sign-up abusals.

Automate the boring stuff is definitely great stuff; I started my web scraping through it. However, I feel it may be a little too much for the 12th grade; again, it may be just me, with very late into the programming interest. My experience when giving recommendations is not to make multiple things learn in parallel. It is always easier to start with a project which we routinely do. The idea is to be clear on the process and outcome, and then achieve it in a non-routine process, programming. The approach will help not easily give up the task, and also brings interest. So, maybe you to list out the tasks that the kids usually do or even ask to list down their favorite/routine tasks and guide them through the programmatic approach.

Be it handwritten or printed text, by far, Microsoft's Azure OCR is best in recognizing text.

Founder - https://extracttable.com here. You should want to give us a try? We are running a closed beta with our premium customers, Happy to provide you access to it.

What metrics do you want the comparison on?

Cost: AWS is not free vs Open Sourced

Time: AWS averages under 10 seconds vs 140 seconds on a standard Dell 7480 & 9 seconds on a GPU Google colab

Character Accuracy: Almost same on a high quality input. No comparison with AWS on a blurred camera photo like this https://github.com/ExtractTable/ExtractTable-py/blob/master/...

Your presumption is either (if you not a h1 candidate) far away from the fact or don't want to accept the truth.

You are looking at the top 100 companies and their full-time employees but not at the 1000s of vendor-employer ponzi scheme/scams that happen every day in the form of contractors who apply with fake resumes, take proxy interviews and outsource the assigned work to someone in India. Doesn't stop there, the dependents of those do a unlicensed food selling, baby sitting, work in local stores, what not. The hard to digest fact is the student who comes on F1 visa, succumbs to the pressure of 90 days period to get into a job to maintain the visa status joins a consultancy company, which starts the game. These consultancies market the candidate with high experience fake resume, arranges proxy candidate to take interviews, even Falsify the requirement of GC/citizenship candidate requirement - all these to charge the client a 100$/hr and pay a mere 40$/hr to the candidate through their sister companies acting as layers. A honest try from USCIS to do a quick scan in containing these will drop at least 50% of H1B lottery applications, void 30% of the consulting company licenses and also help the right talent to get an opportunity. If you are a gov/USCIS rep, drop me an email I'll be more than happy to provide more details, and possible ways of catching the wrong ones.

Hello HN,

I'm Saradhi, taking care of the product and 60% of its development here at ExtractTable. We have launched the service with an HN post in Aug 2019. Today we are 700+ users strong, with over 17000+ free web demo conversions and 145000+ images and PDF pages processed combined through purchased credits.

Although we started to be a developer product, with the request from our premium user, we developed the addon, which I thought will be useful for non-developers and makes it easier for the companies with the workflow integrated to Google Spreadsheets.

Happy to take your feedback and questions.

Thank you.

I agree to this, it's the same with insurance companies too when resolving claims. Feels like they certainly want to make the extraction look complicated for an unknown reason. Not often and not all companies but edge cases

We, extracttable.com, use Google Spreadsheets, a minimal, to save the signup via Google Forms which is also linked to an endpoint using the app scripts to generate an API key.

The biggest workflow I witnessed is from our user, a financial intelligence platform from Brazil, which initially takes data from images and do a manual entry to spreadsheets, has a significant pipeline with as many as 200+ macros defined for their business flow that takes down to close a final output of their clients need.

Prerelease: I planned for a Show HN post, next Tuesday, as we are yet to publish the below content in our site.

We have built a Spreadsheet Add-on (https://gsuite.google.com/marketplace/app/extracttable_image...) to aid their business process that extracts table data from images and puts into the spreadsheet.

Offer: For HN community, take 20% extra credits on your purchase with an email to apikey@extracttable.com with the subject "From HN."

Are you tracking my conversions by IP - to keep me on wait for an hour?

> "getting stuck or even better"

This ain't stop with one. To me, it's just I already lost interest in the current one because my sense felt (or I realized) this is not as worth as what I thought when started. It's okay to move on rather than being stubborn. Since 2017, as a solo developer, I launched 5 projects. Now, most of the time, I'm working on only 2 projects (90-10 split) along side my full time, not a developer.

Also, I was never stuck at an idea phase, I do extensive Google search to validate the idea and readily jump on buying a domain as a commitment to turn idea into the product, with a landing page deployed with in 2 days.

Got it. So, a stat display of users count and api usage ️

We do a lot of this. Honestly, better than what I thought after seeing the title (there are many alike posts). Slick interface. Do not worry about the mobile support - the mobile traffic, not even 3%, does not come for extraction service, they may visit for info, nothing more than that.

From similarweb.com, https://imgur.com/Gd9FG8o, I see the page has over 80% bounce rate which might be the problem?

We are far away from offline mode. May be 2-3 years later thing, if we survive till that time.

Btw, did you check OpenCV morph line detection? It will help to auto detect the evident cell/row boundaries - thus reducing human effort