Swarmfile Blog
Product updates, how it’s built, and what we’re learning from teams running Swarmfile on their biggest projects. For a terser, changelog-style feed of what shipped, see the changelog.
- September 25, 2026
What a read actually costs: measuring Swarmfile on a real 8 GB file
Cold reads of 1 GiB and 8 GiB files on our staging service: bytes moved, time to first byte, the second we were wasting on every cloud read, and the limits we haven't tested yet.
- September 14, 2026
Why our Mac mount doesn't need a kernel extension
What a kext used to cost you on Apple Silicon, what FUSE-T actually is, and the two-detector trick for catching a userspace mount helper that's alive but stopped responding.
- September 12, 2026
Losing a machine shouldn't cost you a shard
Reed-Solomon 10+4, the exact threshold that turns adaptive erasure coding on and off, and how deliberate LAN shard placement uses real rendezvous hashing to guarantee (not just hope) that losing one machine doesn't touch your data.
- September 10, 2026
One office, one download: how the LAN-first P2P actually works
The private QUIC swarm, mDNS peer discovery, the fallback chain from LAN to WAN to cloud, and why there's no explicit "publish" step behind the office's first editor warming the cache for everyone else.
- September 8, 2026
Swarmfile is now a git-LFS server
Point git-LFS at Swarmfile and get real storage, dedup, and quota for your large binary assets, while your source stays in the git host you already use.