Bloom filter utility

Bloom Filter Editor

Configure, train, and export a Bloom filter. Every control explains itself—no external docs required.

Session entries (n) 0

Memory footprint –

Current expected FPP –

Last action Waiting

Settings

Target an error rate, set an expected item count, and override m (bits) or k (hashes) if you want finer control.

Target false positive rate (p) Lower p means a bigger filter but fewer false positives. Expected items (n_expected) How many unique values you plan to insert; used for suggestions.

Suggested size (m)

–

Calculated bits needed to hit the target p for n_expected.

Suggested hashes (k)

–

Number of hash functions that balances speed and accuracy.

Override m (bits) Manual bit count. Must be positive and not absurdly large. Override k (hash count) Manual hash function count. Higher k cuts errors but slows inserts/tests.

Resulting error @ current n

–

Expected false positive probability with the current entries.

Memory footprint

–

Bits, bytes, and KiB/MiB for the current configuration.

Actions

Create a filter, load one from disk, or download your progress as JSON.

Upload JSON Restores a saved filter and metadata.

Loaded file

None

Tracks which JSON is active.

Ready.

Incremental training

Paste newline-separated strings. Empty lines are ignored, duplicates in the same batch are skipped.

Adds all non-empty lines, then clears the box on success.

Usage example

Shows how to reconstruct and query the current filter in client-side code.

// Create or load a Bloom filter to see a tailored example.

Log

Structured activity feed. Mirrors console output.