Introducing Varc: Cursor for Redis

Every Redis GUI shows you a keyspace. None of them show you the traffic — the commands your app is actually sending, in the moment it sends them. When a cache goes wrong at 2am, a tree of keys doesn’t help. You want to hold the next DEL, look at it, and decide.

That’s Varc: a desktop client that treats Redis the way an IDE treats code. Browse and edit at any scale, set a breakpoint on live traffic, step through commands, and rewind writes you didn’t mean to make.

Browse and edit at any scale

Most clients fall over on a big keyspace because they try to KEYS * the world. Varc never materialises the whole keyspace — it renders only the window you’re looking at and pages as you scroll, so a million keys feels like a hundred.

Every value type gets a first-class editor — strings, hashes, lists, sets, and sorted sets — with inline type detection and binary-safe display. Keys and values are bytes, not strings, so a \xac\xed Java-serialised blob shows up as exactly that instead of mojibake.

The part nobody else has: a debugger

Point any client — your app, a script, or Varc itself — at the opt-in local proxy, and every command becomes something you can hold, inspect, and release.

When a command matches, the stream pauses and hands you the held command with its full arguments. Step it, continue, or drop it and synthesise a reply. We’ll go deep on how the proxy stays cluster-transparent in a follow-up post.

It’s fast because it does less

The proxy sits on the hot path, so it can’t be slow. It parses just enough of each RESP frame to route and match, and forwards the bytes untouched when no breakpoint is armed.

Get started in three steps

Varc is pre-release and moving fast. If holding a Redis command mid-flight sounds useful, we’d love for you to try it and tell us what breaks.