Raft Consensus Visualizer
Runs a cluster of 3, 5 or 7 nodes through the Raft consensus protocol, live: followers count down randomized election timeouts (drawn as shrinking rings), time out into candidates, gather RequestVote messages, and the winner starts sending heartbeats. Press "client request" and watch a log entry replicate to followers and commit once a majority has acknowledged it. Then break things: click nodes to kill or revive them, or flip the partition switch to split the cluster and see why the minority side can elect nobody — the majority rule that prevents split-brain. Every node's log is shown as colored cells with a commit marker, message arrows animate between nodes, and an event log narrates each step. Pause, single-step and speed controls let you inspect any moment.
Runs 100% in your browser — nothing you paste leaves your device.
Read the full guide to this tool
Notes
- Randomized election timeouts are Raft's trick against split votes: whoever times out first usually wins before anyone else even becomes a candidate. Watch a re-election after killing the leader.
- An entry is committed once a majority of nodes have it — not when all do. A partitioned minority keeps accepting nothing, and a minority-side "leader" from an old term can never commit, so no split-brain.
- Terms are Raft's logical clock: every election increments the term, and any node seeing a higher term immediately steps down to follower.
- Simplified on purpose: no log snapshotting, no membership changes, and RPCs are single messages with uniform latency rather than batched streams. The core election and replication rules are faithful.
- Runs 100% in your browser — nothing you paste leaves your device.