Full-stack web application to build and deploy computer vision AI models, powered by the getitune library.
Application • Docs • License
Geti is an end-to-end application for building and deploying computer vision AI models. It provides an intuitive graphical interface to upload image or video data, annotate datasets, train and optimize models, and run real-time inference through configurable pipelines.
Main capabilities:
- No-code model lifecycle: move from data upload and annotation to training, evaluation, and deployment in one UI.
- Built-in data and annotation workflows: manage datasets, labels, and revisions with manual and AI-assisted annotation tools.
- Pipeline-based deployment: connect sources (cameras or files) to trained models and route predictions to sinks such as storage, MQTT, or webhooks.
- Edge-oriented optimization: export OpenVINO-optimized models for efficient inference on Intel hardware, with support for other accelerators.
Geti can be installed as a Windows app, run as a Docker container, built from source, or set up with the install script. Pick the option that best fits your environment.
For complete, step-by-step instructions - including prerequisites, GPU/accelerator support, TLS and TURN configuration, air-gapped setup, and troubleshooting - see the Installation guide.
Please check the documentation website for detailed guides, API reference, and other resources to help you get the most out of Geti.
Upgrading an existing installation? See the Upgrade guide for how to move to a newer version (Docker or Windows MSIX) while preserving your projects, datasets and models, with automatic rollback if a migration fails.
Advanced: generate the API spec from source
The OpenAPI specification for the Geti REST API can be generated with the generate-api-spec command:
# From the repo root
cd application/backend
# Generate the OpenAPI spec and save it to a custom location
just gen-api-spec --output-path="openapi.json"The Geti source code is licensed under the Apache 2.0 License. The Windows MSIX App is licensed under the Intel Simplified Software License. For more information, refer to the LICENSE page.
