Install and download

Install CV-Scope

CV-Scope runs on your own computer, with or without a GPU. It needs no account and no cloud service, and the setup downloads two sample videos so the first experiment works without a camera.

Packaged releases

CV-Scope v0.1.0 published 30 September 2026

Download for Windows (64-bit)

CV-Scope-Setup-Windows-x64.exe, 15 MB. Installation notes for Windows

Windows may stop the installer with “Windows protected your PC”, because it is not code-signed yet. Select More info, check that the app is CV-Scope-Setup-Windows-x64.exe, then Run anyway.

SHA-256

f11e32ce44ee8b5e2c23f1e7a81c173d8a713fba1c3efcb63ca35be536d475d6

To check the file before you open it, run this in PowerShell in your Downloads folder. The result must match the SHA-256 above.

(Get-FileHash .\CV-Scope-Setup-Windows-x64.exe -Algorithm SHA256).Hash

Install from the repository

The setup script creates a Python virtual environment, installs the CUDA build of PyTorch when it finds an NVIDIA driver (the CPU build otherwise), installs the frontend, applies the database migrations and downloads the sample videos.

Linux and macOS Windows Docker

            
                git clone https://github.com/ajm2004/CV-Scope.git
                
              
                cd CV-Scope
                
              
                bash scripts/setup.sh
                # .venv, PyTorch, frontend, database, samples
              
                bash scripts/dev.sh
                # API and interface

Needs Python 3.11 or newer and Node 20 or newer; an NVIDIA GPU is optional. Then open http://localhost:5173: the first-run wizard checks the machine, recommends a detector and opens the Scene Builder.

            
                git clone https://github.com/ajm2004/CV-Scope.git
                
              
                cd CV-Scope
                
              
                powershell -ExecutionPolicy Bypass -File scripts\setup.ps1
                
              
                powershell -ExecutionPolicy Bypass -File scripts\dev.ps1

Needs Python 3.11 or newer, Node 20 or newer and Git. Then open http://localhost:5173.

            
                docker compose up
                # CPU image
              
                docker compose --profile gpu up cvscope-gpu
                # NVIDIA image; needs the NVIDIA Container Toolkit

Run from a clone of the repository, then open http://localhost:8420. Data lives in the cvscope-data volume.

Requirements

Python	3.11 or newer (3.11 tested)
Node	20 or newer (for the frontend build)
Operating system	Windows 10/11, Linux, macOS
GPU (optional)	NVIDIA GPU with a driver supporting CUDA 12.6 or newer; Apple silicon (Metal) is supported through PyTorch MPS
Disk	3–5 GB for Python packages with CUDA PyTorch, plus model weights (5–130 MB each)

From docs/installation.md. Development happens on Windows 11 with an NVIDIA GPU; the automated tests run on Linux.

Other ways to run it

CPU only

Everything works on the CPU; the hardware page recommends a smaller detector. Pass --cpu to setup.sh (or -Cpu to setup.ps1) to force the CPU build of PyTorch.

Manual CPU setup

NVIDIA GPU

Needs a driver that supports CUDA 12.6 or newer. The Hardware page shows whether PyTorch can use the GPU.

CUDA setup and troubleshooting

Apple silicon

The default PyTorch wheels provide Metal (MPS) acceleration; choose device auto or mps.

Apple silicon notes

Docker and PostgreSQL

CPU and NVIDIA images, with an optional PostgreSQL service. Data lives in a named volume.

Docker setup

Permissive licences only

Skip the Ultralytics extra (AGPL-3.0) and use the BSD-3-Clause torchvision detectors or your own ONNX models.

Install without Ultralytics

Opening it to a network

CV-Scope has no login and binds to 127.0.0.1. Put it behind a reverse proxy with authentication and TLS first.

Run a real study

Check the installation

With the virtual environment active, the command line shows what CV-Scope found and measures this machine.

cvscope hardware            # hardware probe and recommendations
cvscope models list         # catalog with install state
cvscope models install yolo11n
cvscope benchmark yolo11n --video samples/videos/people-detection.mp4
python -m pytest backend/tests

Release history

CV-Scope v0.1.0 30 September 2026