Doesn't work for the <i>other</i> 125 encoded characters that are numerically seven, as defined in the Unicode Character Database:<p><a href="https://www.unicode.org/Public/17.0.0/ucd/extracted/DerivedNumericValues.txt#:~:text=DIGIT%20SEVEN" rel="nofollow">https://www.unicode.org/Public/17.0.0/ucd/extracted/DerivedN...</a><p>(Viewable / copy-able version: <a href="https://pastebin.com/fNRv3wD6" rel="nofollow">https://pastebin.com/fNRv3wD6</a>)
Heads up OP: I’m trying to get a pro license but your checkout flow is borked. Probably should fix ASAP before missing out on HN front page traffic.
I was almost sure that they do it client-side for a joke, but if you check the browser console, you can see that they actually make a request! You can even make the same request with curl and it works!<p>Although, making an HTTP request manually is quite inconvenient, so I'm waiting for Python SDK.
Vulnerable to a distillation attack, unfortunately not much of a moat.
I misread this as "is even" and was shocked that 46 returned false
7.0000000000000001 evaluates to true.
Bug report: I tried 6.999999̅ and got false. So there's some nonstandard model of the reals being leveraged here.
> 77.7% uptime SLA<p>Giving GitHub a run for its money, I see.
That documentation is woefully inadequate. It provides only one example request, and then it shows two separate responses, and it doesn't make clear which one is associated with the request. It doesn't even describe the individual request fields, nor does it provide any response codes or a list of error codes/messages. How am I supposed to develop with this?<p>I do, however, appreciate the seven figure SLA. My service requires at least five nines of uptime, and seven figures is definitely more than five.
Bug report: I entered 3 + 4 and did not get a kinda or true.<p>This app is ngmi
Well the all only has 77.7% uptime maybe it just returns wrong things while it’s down, that’s probably it. Try upgrading to Enterprise that’ll probably fix it.
I would have expected something other than false for "se7en".
Similar report here:<p>70/10, 7.1-0.1 and srqt(49) also do not return true.<p>Is there a published SLA for the free version?
00000111 also came back false
no Roman numeral support either
BTW: For a tool that actually legitimately does this, look at Semgrep. Their playground example literally assigns 1 to a variable x, after which searching for "2" finds the expression "1 + x" in the code: <a href="https://semgrep.dev/playground/s/5rKgj" rel="nofollow">https://semgrep.dev/playground/s/5rKgj</a>
complaint: someone entered "seven" and it crashed my entire infrastructure because the output returned a non standard 'kinda'.
Reminds me of <a href="https://five.js.org/" rel="nofollow">https://five.js.org/</a>
Enterprise looks promising, but before I take this to upper management: How many sevens of uptime are we talking?
Needs an Agent skill! Gotta be more modern :)
Apparently “seven” is only kinda seven. I would argue that seven is seven!
I can use this as a random number generator; at least it's not nine.
It hallucinates 6.9999999999999996 to be seven.
I wanted to subscribe and I can't! How do you expect to make any money if that doesn't work?!!1
No other numbers were harmed in the making of this API.<p>But their feeling hurts, especially primes.
This Is Seven as a Service.
This SaaS actually will be replaced with an in house vibecoded solution.
Is this SOC2 compliant?
We wanted to subscribe to the enterprise plan, but unfortunately:<p>- No Soc-2 compliance<p>- No sso support.<p>We asked if we could host on-prem or even byoc but that seems an impossible dream.<p>Smh
TIL 6+1 is not seven.
Does not work for Nw==
Does this have an MCP server?
I will vibe code my way out of poverty:
I like how you spent $10 for the domain for this. :)
Eh, more better than Prolog.