SmyteStreamProcessingArchitecture

Smyte Stream Processing in 7 Steps

Smyte stream processes features extracted from actions (similar to events), interprets custom rules, makes sophisticated computations in parallel, and returns verdicts in near real-time. What makes Smyte unique is our fraud and abuse classifiers that are transparent, fully customizable, and based on live streaming data.

Read more…

High-Performance Rate Limiting


At Smyte we spend a lot of time stopping spam, scams, credit card fraud and online harassment. We’re open-sourcing one of the fundamental tools we use to stop malicious user behavior: our high-performance rate limiter.

Read more…

Counting with Domain Specific Databases

Counting is core to what we do at Smyte. Our realtime spam and fraud fighting software analyzes an event stream and needs to answer many questions about it. Effective questions often begin with “how often” or “how many”.

Read more…

Streaming Similarity Search for Fraud Detection

Bad actors on the internet often pretend to be who they are not. But if one looks closely at their actions, one can see who they really are. At Smyte, we seek to identify who those bad actors are and stop them in their tracks before they can do you or your company significant harm. In this brief post, we explore one of those techniques: similarity search.

Read more…

Announcing jsxstyle 1.0

jsxstyle is now 1.0. We’ve been using it for years at Smyte, and we’ve learned a lot about what works and what doesn’t when it comes to inline styles.

Read more…

Subscribe Here

Zendesk OnDemand Webinar

Recent Posts