Skip to content
1AIVault1AIVault
Recall

Full-text search with ranking

Ranked keyword search for when you know the exact term you're hunting.

Overview

What Full-text search with ranking does

Full-text search covers the cases where you know precisely what you want: a function name, an error string, a client, a product code. It matches literal terms across entry titles, source labels, and the full memory body, then orders results with BM25 ranking so the strongest matches rise to the top instead of arriving in raw date order.

Keyword and semantic search complement each other. Semantic recall is best for fuzzy, half-remembered ideas; ranked full-text search is best for auditing exact strings and confirming whether something specific was ever discussed. Because it ranks rather than merely filters, a query for a rare identifier lands its handful of real hits first, which keeps precise lookups fast even in a vault of thousands of entries.

FTSBM25
Why it matters

The payoff for your AI memory

Exact-term precision

Search for a specific name, error, or identifier and get the entries that literally contain it, not approximate neighbours.

Relevance-ranked results

BM25 scoring pushes the strongest matches to the top, so the most relevant hit is usually the first one you see.

Auditable by design

Because matches are literal, full-text search is ideal for confirming whether an exact decision, term, or source ever appeared in the vault.

How it works

From first launch to reusable memory

  1. 1

    Type the exact term, phrase, or identifier you remember into vault search.

  2. 2

    1AIVault matches it across entry titles, source labels, and memory bodies.

  3. 3

    Read results in BM25 relevance order, strongest match first.

  4. 4

    Combine with source or category filters to narrow a large result set.

FAQ

Common questions

What is BM25 ranking?

BM25 is a proven relevance-scoring method that weighs how well each entry matches your query terms, so the most relevant memories appear before weaker, incidental matches.

When should I use full-text search instead of semantic search?

Use full-text search when you know the exact wording, such as an error string or identifier. Use semantic search when you only remember the idea, not the words.

Local memory, shared everywhere

Give every AI tool the same memory.

Start free, import real conversations, and reuse your memory across every AI agent you already use.