Count 1 billion unique visitors in 12 KB: HyperLogLog steps in wpipe.

Counting unique visitors shouldn't bankrupt your Redis RAM or choke your SQL database. wpipe-steps gives you native Redis HyperLogLog steps to estimate billions of distinct items in 12 KB.

Here is how you use Redis HyperLogLog Cardinality Estimation in a production pipeline with wpipe-steps:

from wpipe import Pipeline
from wpipe_steps.database.redis.hyperloglog import (
    redis_hll_add_sync,
    redis_hll_count_sync
)

pipeline = Pipeline(pipeline_name="unique_traffic_counter")

pipeline.set_steps([
    # Add unique visitor IP addresses
    redis_hll_add_sync.as_step(
        name="record_visitor",
        key="traffic:unique_ips:2026-09",
        elements=["192.168.1.1", "10.0.0.42", "172.16.0.99"]
    ),
    # Retrieve approximate cardinality in O(1) time
    redis_hll_count_sync.as_step(
        name="get_unique_count",
        keys=["traffic:unique_ips:2026-09"],
        response_key="unique_visitors"
    )
])

result = pipeline.run({})
print(f"Estimated unique visitors: {result['unique_visitors']}")

Why developers love wpipe-steps:

• 196 cataloged steps covering Redis, ClickHouse, MySQL, WAF, S3, Docker, and local HuggingFace AI.
• Lazy-loading imports for instant sub-100ms startup times.
• Clean .as_step() factory interface.

Check out the full repository on GitHub!