Answers before you connect.

Varc is a Redis and Valkey client with a built-in debugger. Browse and edit keyspaces as usual, then opt into a local proxy when you need to inspect live traffic, set breakpoints, step commands, or rewind a write.

Both. The desktop app is a full client for browsing and editing data, with the debugger layered on top as an opt-in workflow. You can use the client normally and only route traffic through the proxy when you need to investigate behavior.

Most Redis GUIs show you the current state of a database. Varc is also designed to show how that state changes: the local proxy can hold commands in flight, and captured before-images make an exact rewind possible.

Varc is being built for macOS, Windows, and Linux. The desktop app and headless CLI share the same Rust engine so the core connection, debugger, and rewind behavior can stay consistent across platforms.

Point your app, or Varc itself, at Varc's local proxy. Commands pass through normally until a breakpoint matches. Varc can then hold the command so you can inspect it, step to the next command, continue, skip it, or release it.

Breakpoints can target any command, writes only, a specific command name, a key pattern, or a single connection. The same controls work when a command is part of a pipelined batch.

Yes. A pipelined batch can be held and stepped one command at a time, so you can see which operation changes the state instead of treating the whole batch as one opaque event.

Held commands are dropped on client disconnect. Varc does not execute them a second time after the connection is gone.

Varc captures a before-image before a write changes the key. Rewind uses that captured value and expiry to restore what was actually there, rather than trying to guess a reverse command later.

Yes. The before-image includes the expiry state, so restoring a key can preserve its TTL rather than bringing back only the value.

Yes. Varc is designed for Redis and Valkey, with the same client model across those Redis-compatible systems.

Standalone, Cluster, and Sentinel are first-class. Browsing, debugging, undo, scripts, and load-test workflows use the same topology-aware routing layer.

Varc uses windowed views instead of materializing the whole keyspace in the UI. That keeps browsing useful as the dataset grows and avoids turning a scan into a memory problem on the client machine.

Yes. Varc handles keys and values as raw bytes end to end. The client does not silently convert binary data to text while you browse, edit, or export it.

Yes. SSH host-key verification, TLS, and mutual TLS are treated as first-class connection fields. The local debugger proxy does not require you to weaken the transport it is inspecting.

Yes. varc-cli uses the same Rust engine as the desktop app, so you can run the proxy, replay a session, or step a breakpoint from a terminal or CI job.

The headless CLI is designed for terminal and CI workflows. It can run debugger and replay operations without requiring the desktop window.

Yes. Sandboxed JavaScript transformers can reshape values in the viewer and inspector. The runtime applies hard memory and time limits so a decoder cannot take over the client.

Yes. Free connects directly to Redis or Valkey without an account. It includes one saved connection, one active connection, the key browser, search, editing, and server information.

No. Personal is a paid perpetual licence. A purchase includes one year of new Varc builds, and every build released during that update period remains yours to use. Renew when you want newer builds.

Personal is $99 one-time, Pro is $249 one-time, and Team is $399 per seat with a two-seat minimum. Each purchase includes one year of new builds; covered builds remain usable forever.

Bring your own key. Personal, Pro, and Team can connect a supported AI provider using credentials you control. Pro and Team also include a Varc-hosted usage allowance.

Every plan supports Redis and Valkey across standalone, Sentinel, and Cluster deployments, including SSH, TLS, and mTLS connections.

Cloud KV is a Pro and Team capability with a 50 MB hard cap per database. Pro currently includes three databases; the Team database allowance is still to be decided.