FLAT Index

Compares every stored vector one by one — 100% accurate, exhaustive search

FLAT checks every stored vector one by one. Exact but slow with large datasets.

Flow: 📥 Query→🔄 Compare All→📊 Rank by Score→🏆 Top Match

🎯

Query Vector

Editable

x y z

🗃️

Edit any value. FLAT compares the query against all of these.

🔢

Comparisons Made

4 vectors checked

FLAT always checks every stored vector, no shortcuts

📊

Live

Full scan, no shortcuts. Guarantees the true nearest neighbor every time.

Small datasets, exact accuracy, benchmarks, prototyping.

🎓 Tips 3 tips

Always exact. FLAT compares the query against every stored vector, so it never misses the true nearest neighbor.

Add more vectors and notice the comparisons count grows one-to-one, FLAT has no shortcuts, so search time scales linearly with dataset size.

FLAT is ideal for small datasets or when 100% accuracy matters more than raw speed.

💡

Key Takeaway

FLAT guarantees the exact nearest neighbor by checking every vector, accurate but slow as the dataset grows.