LOCAL-FIRST · NO SAAS SHORTCUTS

A tracker I built to prove a point.

Every mainstream tracker leans on a hosted sync provider I don’t control. So I wrote my own from scratch: optimistic mutations, a server-authoritative log, conflict-free merges, an offline queue that survives reload. Then I put a Linear-shaped tracker on top so you can actually feel it work.

Try the demoOr sign up

The demo workspace is shared on purpose. Open it in two tabs, edit both, watch them converge live. That’s the sync engine, doing the thing.

HOW THE SYNC ENGINE WORKS

Write locally. Reconcile in the slipstream.

Every edit lands in a local log the instant you make it. No spinner, no round trip. When the network shows up, a CRDT folds concurrent edits together, so two people working offline still end up on one truth. You don’t notice it. That’s the compliment.

  • Edit hits the local write-ahead log immediately.
  • Deltas stream to peers as bandwidth allows.
  • CRDT merge guarantees every replica converges.
youpeerCRDTmerge✓

WHO BUILT THIS

I’m Hiten.

I’m a full-stack engineer. I built Slipstream partly because I wanted to go deep on the mechanics of a sync engine (mutators, optimistic apply, deterministic rebase, transactional counters, CRDTs, a WebSocket poke channel) and partly because I got tired of every demo project reaching for the same three SaaS logos. If you’re curious about the engine, everything is open-source, the commit history reads as a build log, and the architecture doc lives in the repo.

source on GitHubarchitecture dochiten.dev