10x faster
than PostgreSQL
PostgreSQL extension that uses machine learning to predict data locations, eliminating tree traversal for sequential workloads.
Real Performance Numbers
Tested on real datasets, not synthetic benchmarks
| Dataset Size | B-tree | Learned Index | Speedup |
|---|---|---|---|
| 10,000 | 3.5M ops/sec | 8.1M ops/sec | 2.3x |
| 50,000 | 2.7M ops/sec | 6.5M ops/sec | 2.4x |
| 100,000 | 3.2M ops/sec | 8.0M ops/sec | 2.5x |
| 500,000 | 2.6M ops/sec | 7.2M ops/sec | 2.8x |
Range queries see up to 16x improvement
How It Works
Replace 45-year-old B-trees with machine learning
Traditional B-Trees
×
Each level requires a disk read ×
Each read is a CPU cache miss ×
200+ nanoseconds per lookup ×
O(log n) complexity forever Learned Indexes
✓
Analyze data distribution (~100ms) ✓
Find location in 1-2 CPU instructions ✓
Binary search ±100 positions ✓
O(1) prediction + small refinement Try It Now
Multiple ways to experience 10x faster database performance
PostgreSQL Extension
Add learned indexes to your existing PostgreSQL database
# Install our extension
CREATE EXTENSION omendb;
# Benchmark on your data
SELECT learned_index_benchmark(10000);
Advanced Usage
Optimize your PostgreSQL queries with learned indexes
# Create indexes on sequential data
CREATE INDEX ON events (timestamp);
# Benchmark your queries
SELECT learned_index_benchmark(50000);
Why This Matters
Every millisecond of latency costs money
💰
E-commerce
100ms delay = 1% lost sales
📈
Financial Trading
1ms advantage = millions in profit
⚡
Real-time Analytics
Faster queries = better decisions
Significant performance improvements for time-series and sequential data workloads.
Try the Extension
PostgreSQL extension available now on GitHub
Open source and ready to use