Documentation
Overview
Modern operating systems were designed 30 years ago around files and folders. Today, knowledge workers live across 40+ browser tabs, multiple IDEs, local files, terminal sessions, and messaging apps simultaneously.
KNEMOS is a local-first AI productivity system that acts as a cognitive layer between the user and their computer. It automatically clusters your entire digital workspace into intelligent semantic groups, makes your screen history searchable in natural language, and measures your cognitive performance using Wolfram Language analytics.
Step-by-Step Guide (For Beginners)
Follow these instructions to get your KNEMOS productivity system up and running in minutes, even if you don't know how to code.
Download & Install the Desktop App
Sign up or Log in to this website. Once authenticated, navigate to the Downloads section and download `KNEMOS-Setup.exe`. Double-click the file to install it on your Windows computer. This will install both the visible UI and the invisible AI brain.
Open KNEMOS & Let it Boot
Launch KNEMOS from your Start Menu. The very first time it boots, it may take a few minutes to silently install the local AI models (Tesseract OCR & ChromaDB vectors) on your computer. You will see a dark loading screen. Do not close it.
Copy your Secret Authentication Token
Once the dashboard opens, navigate to the Settings panel (the gear icon). Scroll to the bottom and find the "Auth Token" section. Click the button to copy your unique, randomly-generated secure token to your clipboard.
Install the Chrome Extension
To allow KNEMOS to track what browser tabs you use (so it can calculate your Focus Score), install the KNEMOS Extension in Google Chrome. Open the extension, paste the token you copied in Step 3, and click "Connect".
Start Focusing!
You are all set! Group your tabs and apps into a "Workspace" in the desktop app. Click "Activate Deep Focus" to instantly blackout distractions. Any apps not in your workspace will automatically be minimized to keep you on track!
Core Features
Semantic Workspace Clustering
AI automatically groups your browser tabs, VS Code windows, terminal sessions, and local folders into named semantic workspaces—no manual tagging or folder creation required.
Memory Lane
Periodically captures screenshots, runs OCR, generates embeddings, and indexes everything into ChromaDB. Search your entire workspace history in natural language.
Wolfram Intelligence Layer
Computational analytics providing deep productivity forecasts, context-switch tracking, and memory relationship graphs natively via Wolfram Engine.
RAM Recovery Engine
Intelligently hibernates inactive workspaces and calculates live RAM/CPU savings, tracking efficiency through the local backend.
AI Pipeline
psutil + pywin32 + watchdog + mss + Chrome Extension
mxbai-embed-large (via Ollama). High-fidelity semantic vectors.
HDBSCAN. Semantically related resources → workspace clusters.
Ollama + Qwen2.5-7B (standard) / Qwen2.5-3B (low-end).
Tesseract OCR + ChromaDB. Searchable vector memory.
Wolfram Language. Cognitive Focus Score + predictions.
Technology Stack
| Desktop Shell | Tauri v2 (Rust-native backend) |
|---|---|
| AI Backend | FastAPI, Python 3.11, APScheduler, WebSockets |
| AI / ML | mxbai-embed-large, HDBSCAN, Ollama + Qwen2.5 |
| Vector DB & OCR | ChromaDB, Tesseract |
| System Monitor | pywin32, psutil, watchdog, mss |
| Auth & Web | Supabase Auth, Next.js 15, TailwindCSS |
Privacy First
No screenshots, embeddings, or workspace data are ever transmitted externally. The cloud only handles authentication and app updates. Everything else runs purely on 127.0.0.1.