Skip to content
1AIVault1AIVault
Classify

Smart Topic Classification

Point 1AIVault at a local Ollama model and it extracts topics from your entries, merges duplicates, and groups related memories — cross-classification across topics is the default.

Topics view with digest and timeline.
Overview

What Smart Topic Classification does

Point 1AIVault at a local Ollama model and it extracts topics from your entries, merges duplicates, and groups related memories — cross-classification across topics is the default.

A growing vault is only useful if you can find related context fast. Topic classification turns a flat list of entries into a navigable map of subjects, projects, and decisions — without sending anything to a remote service.

TopicsClassification
Why it matters

The payoff for your AI memory

Ollama-powered extraction with `qwen2.5:7b` as the default

Ollama-powered extraction with `qwen2.5:7b` as the default model

Suggests topic merges with a side-by-side confirm/reject

Suggests topic merges with a side-by-side confirm/reject dialog

Cross-classification — an entry about Postgres for

Cross-classification — an entry about Postgres for the analytics dashboard appears under both topics

Auto-generated topic digests so you can catch

Auto-generated topic digests so you can catch up at a glance

How it works

From first launch to reusable memory

  1. 1

    Import conversations so raw memories flow into the vault.

  2. 2

    Let local classification group entries into topics automatically, or run a deeper pass with a connected AI client.

  3. 3

    Browse the topic list to see every area of work at a glance.

  4. 4

    Rename, merge, or refine topics until the map fits your mental model.

FAQ

Common questions

Do I need Ollama installed?

Yes for the classifier. If Ollama isn't running, classification is disabled and an inline banner in the Dashboard prompts you to set it up. The rest of the vault works without it.

Can I use a different model?

Yes. Settings → Classification lets you pick any Ollama model you have locally. `qwen2.5:7b` is the default because it balances quality and speed on most laptops.

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.