4 ms·

Smyte (YC W15) Launches to Fight Online Fraud

65 points by peterhunt 11y ago

peterhunt 11y ago
Co-founder here -- happy to answer any questions!
- sgrove 11y ago
  I'm curious what data I have to send to you to get this to work - would flipping a Segment.io switch be enough to get started?
  - peterhunt 11y ago
    Hey Sean! Our developer docs at http://docs.smyte.com/ http://docs.smyte.com/ should have the answers here, but basically you can drop a script tag on your page to get a basic view of what's going on, and use our REST API server-side to annotate that basic view with additional data. Ideally, you'd serialize most of your models to JSON and send them over to Smyte so it can detect patterns and relationships. Segment is a good idea; I should chat with them.
  - qixxiq 11y ago
    Josh from Smyte here. Definitely check the docs as Pete suggested, but would also like to point out that we'll take just as much data as you're comfortable with and/or can send for privacy reasons. We have a ton of different signals that make our system tick, but based on whatever you send us it'll function as best as it can.
- rebelidealist 11y ago
  What are some ways you are different from Sift Science?
brandonb 11y ago
I think I might be your competitor from eight YC batches ago, but good luck! You seem to have a great team.
- peterhunt 11y ago
  Thanks! Based on my understanding of Sift Science as an outsider, we're focused on a different market. We're explicitly trying to stay away from traditional e-commerce applications (shopping carts, standard transactional businesses, etc) because the space is so crowded, preferring instead to focus on trust and safety applications (i.e. two untrusted users interacting with each other on some service vs 1 untrusted user interacting with a merchant).
JoeAltmaier 11y ago
Great summary - there's lots of subtleties in this space and most of our home-brew solutions would miss a lot. This company is an example of experts leveraging their knowledge to benefit us all. Well done!
Chronic31 11y ago
Just another fraud detection company.
tomasien 11y ago
This looks great! My last company would have loved this.
jaz46 11y ago
Congrats guys!

65 points by peterhunt 11y ago

peterhunt 11y ago

Co-founder here -- happy to answer any questions!

sgrove 11y ago
I'm curious what data I have to send to you to get this to work - would flipping a Segment.io switch be enough to get started?
- peterhunt 11y ago
  Hey Sean! Our developer docs at http://docs.smyte.com/ http://docs.smyte.com/ should have the answers here, but basically you can drop a script tag on your page to get a basic view of what's going on, and use our REST API server-side to annotate that basic view with additional data. Ideally, you'd serialize most of your models to JSON and send them over to Smyte so it can detect patterns and relationships. Segment is a good idea; I should chat with them.
- qixxiq 11y ago
  Josh from Smyte here. Definitely check the docs as Pete suggested, but would also like to point out that we'll take just as much data as you're comfortable with and/or can send for privacy reasons. We have a ton of different signals that make our system tick, but based on whatever you send us it'll function as best as it can.
rebelidealist 11y ago
What are some ways you are different from Sift Science?

brandonb 11y ago

I think I might be your competitor from eight YC batches ago, but good luck! You seem to have a great team.

peterhunt 11y ago
Thanks! Based on my understanding of Sift Science as an outsider, we're focused on a different market. We're explicitly trying to stay away from traditional e-commerce applications (shopping carts, standard transactional businesses, etc) because the space is so crowded, preferring instead to focus on trust and safety applications (i.e. two untrusted users interacting with each other on some service vs 1 untrusted user interacting with a merchant).

JoeAltmaier 11y ago

Great summary - there's lots of subtleties in this space and most of our home-brew solutions would miss a lot. This company is an example of experts leveraging their knowledge to benefit us all. Well done!

Chronic31 11y ago

Just another fraud detection company.

tomasien 11y ago

This looks great! My last company would have loved this.

jaz46 11y ago

Congrats guys!