EN | TR
Sign in Try it free
Semantic Search

Your Knowledge,
Understood — Not Just Searched.

TotalApp's AI-powered tools can find and surface relevant material from your personal knowledge library automatically — ranked by meaning, not keywords, and running entirely on your own machine.

Try Semantic Search See how it works
Local
Runs on your machine via Ollama
Live
Re-ranks as you type
2
Search modes: Assisted & Automatic
0
Data ever leaves your device
How It Works

Ranked by meaning, not keywords

Semantic Search converts your knowledge library and your current input into text embeddings, then compares them directly using cosine similarity — surfacing the most relevant material first, without you ever typing a search query.

  • ✓
    No manual search box — ranking happens automatically as you type
  • ✓
    A relevance score badge shows how closely each source matches
  • ✓
    Lightweight ranking layer — not a full RAG pipeline or vector database
  • ✓
    Works instantly for personal libraries of tens to low hundreds of sources
See It in Action
totalapp.app/essay-writer
Attach Knowledge
Ranked by relevance to your topic
"Climate policy in emerging markets"
📄 Emerging Markets Climate Report 2026 71%
📄 COP Policy Briefing Notes 64%
📄 Q3 Marketing Plan 28%
📄 Team Offsite Notes 21%
Writer Tools Example

The first place it's wired up

Today, every AI Writer screen — Blog Writer, Essay Writer, Article Writer, Storyteller, Anime Writer, Scenario Writer, and Smart Rewriter — uses Semantic Search behind its Attach Knowledge button. It's simply the first place this capability was built; more AI-powered tools are expected to use it over time.

  • ✓
    Attach Knowledge button ranks your library as you type a topic
  • ✓
    Ground generated content in real documents instead of guesswork
  • ✓
    Cite, reference, or build on material you've already written
  • ✓
    Works the same way across every Writer Tool screen
Explore Writer Tools
totalapp.app/blog-writer
Blog Writer
Insert Prompt Attach Knowledge
Sustainable Packaging in 2026
Grounded in 2 attached sources
Based on your Q3 Sustainability Report and Supplier Materials Brief, sustainable packaging adoption has accelerated across three key verticals
Search Modes

Assisted or fully automatic

Choose how much control you want. Assisted mode ranks sources but leaves the final pick to you. Automatic mode attaches the best matches for you — no clicks required.

  • ✓
    Assisted (default): ranked suggestions, you confirm the attachment
  • ✓
    Automatic: top matches auto-attach as you type, with an override anytime
  • ✓
    Choose raw cosine similarity or a normalized 0-100% score display
  • ✓
    Configured once in Settings, applies everywhere Semantic Search runs
Configure Search Modes
totalapp.app/settings/agentic
Embedding Engine
My Knowledge Search Behavior
Assisted — I pick from ranked suggestions
Relevance Score Display
Cosine similarity (raw, default)
Embedding Model
nomic-embed-text via Ollama
Local & Private

Your library never leaves your machine

Embeddings are computed by a local Ollama instance running on your own computer. TotalApp's cloud server has no network path to reach it — ranking always happens directly between your browser and your own machine.

  • ✓
    Powered by Ollama, installed and run entirely on your device
  • ✓
    Independent of your Writer Engine (API, Local CLI, Local LLM, Ollama)
  • ✓
    No embedding model configured? Everything still works, just unranked
  • ✓
    No subscription or extra cost — bring your own local model
See All AI Features
127.0.0.1:11434
Local Ollama Instance
Model: nomic-embed-text — running on your machine
My Knowledge Library
embed · local
Embedded
↓
Current Topic
embed · local
Embedded
↓
Cosine Similarity
compare · local
Ranked

Find what matters, automatically.

Semantic Search is built into TotalApp's AI tools today — starting with the AI Writer suite — and designed to expand as more AI-powered features ship.

Get Started Free Read the Docs