Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
1–10 of 11 posts
Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#2Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#3Your security page has a little typo: "Veryfi’s is machines all the way down."
Wanted to make sure I let you know!
Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#4Super cool, love the landing page — though not yet able to afford to offer this to my clients in my app anytime soon. Your security page has a little typo: "Veryfi’s is machines all the way down." Wanted to make sure I let you know!
Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#5Congrats on your launch and recent 12M funding. I'm curious about your business - we are in a similar (not the same) space. Your pricing of $0.08 Per Receipt, $0.16 Per Invoice is a big surprise. What is your acquisition channel?
Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#6Congrats on your launch and recent 12M funding. I'm curious about your business - we are in a similar (not the same) space. Your pricing of $0.08 Per Receipt, $0.16 Per Invoice is a big surprise. What is your acquisition channel?
Unrelated to your question, but as someone not familiar with the space, is that pricing competitive?
Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#7This looks like something I'd definitely use (and pay for, at that price).
Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#8Congrats on your launch and recent 12M funding. I'm curious about your business - we are in a similar (not the same) space. Your pricing of $0.08 Per Receipt, $0.16 Per Invoice is a big surprise. What is your acquisition channel?
Unrelated to your question, but as someone not familiar with the space, is that pricing competitive?
Alternative solutions are things like Mechanical turk at 1c per field or other solutions which vary in accuracy and typically use human in the loop and can range up to $1 per invoice.
In the end, it all depends on your application. If you need 1 or 2 fields then you can get away with other solutions. If you are building a competitive secure product and will use this wealth of data for business intelligence (cpg retail, bill pay, hotel folio) then 16c for a starter plan without a contract is a great price.
There's a free 14 day trial here https://hub.veryfi.com/api/hn/ and just requires your email to experience the product. Nothing to loose.
PS. I'm one of the founders.
Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#9How accurate is this? I tried ocr'ing my receipts a while back and the issues with misrecognized numbers were too large to be worth it. Unlike text where a few typos are unlikely to largely change the meaning, a single digit wrong on a receipt can be a big issue and receipt numbers aren't super high fidelity. This looks like something I'd definitely use (and pay for, at that price).
Let me know what you think.
PS. I'm one of the founders.
Re: Show HN: OCR Invoices and Receipts in Seconds Using Go SDK
#10How accurate is this? I tried ocr'ing my receipts a while back and the issues with misrecognized numbers were too large to be worth it. Unlike text where a few typos are unlikely to largely change the meaning, a single digit wrong on a receipt can be a big issue and receipt numbers aren't super high fidelity. This looks like something I'd definitely use (and pay for, at that price).
You can Test Drive It here https://www.veryfi.com/data-intelligence/ (halfway down the page) without signup or a 14 day free trial here https://hub.veryfi.com/api/hn/ (just email req to get going) Let me know what you think. PS. I'm one of the founders.
I took a photo of a slightly curled but not wrinkled receipt on a wood desk. The receipt had dark printing on the back which partially showed through slightly in areas (though not behind/near the prices).
The item prices were correct, as was the total. It confused ¥79 as 179 for tax. The item names (Japanese, katakana) were wrong but mostly right, and seemed wrong in consistent ways rather than randomly. It said the store was Target, but it was actually a local store chain with an unrelated name/logo.
I'm going try it on a couple more if I have a chance.