Velaxe
Knowledge Base / FAQ — Multi-locale Articles, Attachments & AI Search | Velaxe

Knowledge Base / FAQ

Tune search: FTS + semantic vectors for better answers

Use keywords + embeddings, analyze logs, and improve ranking.

18 min Intermediate Support Lead, Docs Writer Updated Jul 30, 2025

Overview

Use keywords + embeddings, analyze logs, and improve ranking.

Prerequisites

None.

Permissions required

kb.view kb.manage

Downloads & Templates

Steps (3)

Estimated: 18 min
  1. 1

    Review search logs

    Support Lead 5 min Back to top

    Open Settings → Telemetry → Search Logs. Identify top queries with low CTR.

    Tips

    Validation

    Success criteria

  2. 2

    Improve titles & summaries

    Docs Writer 8 min Back to top

    Refine article titles to match query phrasing; add short first-paragraph summaries.

    Tips

    Validation

    • CTR improves on re-tested queries.

    Success criteria

  3. 3

    Re-run semantic search

    Support Lead 5 min Back to top

    Trigger a re-index (if configured) and verify top-k results.

    Tips

    Validation

    Success criteria

About this guide

Knowledge Base / FAQ is an enterprise-grade knowledge system for teams to author rich articles, attach images/videos/files, tag them for easy discovery, and localize content by locale. Users get instant answers via hybrid search: classic full-text (FTS) plus semantic vector retrieval with embeddings.

Editors benefit from autosave, clean Markdown/HTML authoring, tag chips, and attachment management. Ops can choose the embedding model and enable a nightly drift job that re-embeds older content as models improve—keeping search results fresh and accurate.

The app is workspace-native, so your data lives with your workspace, behind Velaxe RBAC and audit logs.