Log Clusterer - Group Repeated Log Patterns

Log Clusterer groups repeated log patterns to help you find signal in noisy log files. Raw server logs often contain thousands of identical lines from retries, health checks, or background tasks — these drown out the interesting events that actually indicate problems or behavior. This tool uses TF-IDF vectorization and DBSCAN clustering to group similar log lines together, so you can quickly see which patterns occur most frequently and which are anomalies.

Each log line is tokenized and converted to a TF-IDF vector (term frequency-inverse document frequency). DBSCAN clusters the vectors in high-dimensional space, using auto-estimated epsilon based on k-nearest-neighbour distances. Lines that don't fit any cluster become singletons (anomalies). Each cluster shows a representative pattern, occurrence count, and sample lines.