By popular request, today we released **Atomdrift Scan v2.9.0, which now ships with a stable HTTP API. But that's not all! We also are announcing Beamline, our end-user low-latency API surface, which runs as a Cloudflare Worker.
Why go to all this trouble? Because users wanted something more than just a CLI tool.
This meant a couple of things had to happen:
- We had to plumb bearer tokens in anywhere
- We had to design a friendly JSON API
- We had to lower latency everywhere to make it reach our goal of <4s uncached analysis
So why the extra Beamline layer? Edge caching. Cloudflare Workers are embarassingly cheap to run, particularly if you use their NodeJS workers and cache. The idea is: you point end-users to Beamline; and then you run scan servers wherever you want. Beamline will handle the load-balancing and caching, ensuring most responses arrive back in <50ms.
We also upgraded our other tools to support the new scan release:
cleave v2.9.0 preserves trait suppression and interdependency data for modelling malecules and providing additional context to LLMs.
filefacts v1.5.0 offers better improved PE data, especially for handling expired certs.
stng v1.10.0 now recoves more XOR-encoded data.
fletch v1.2.0 offers improved PURL handling, particularly for PURLs that ship with multiple files underneath them.
Release notes: scan v2.9.0 · cleave v2.9.0 · stng v1.10.0 · filefacts v1.5.0 · fletch v1.2.0
brew upgrade atomdrift-project/tap/cleave atomdrift-project/tap/scan atomdrift-project/tap/stng