Documentation

What is Git AI?

Git AI, emphasizing the importance of avoiding vendor lock-in, streamlines AI development by replacing a pool of tools with a single interface for building and deploying AI. This MLOps platform allows customers to fast-track their AI deployment, scale effectively, and maintain ownership of their AI across any cloud environment, whether public or private. Featuring a Git-native web app and command line interface, Git AI simplifies the process to build, run, and track experiments through a unified configuration. This approach ensures users have the flexibility to move their operations without being bound to a single provider, fostering innovation and adaptability.

Why choose Git AI?

Choosing CodeDepot's Git AI leverages the familiarity and widespread use of Git among developers, offering a seamless transition into AI development and deployment. With Git being integral to over 95% of developers' workflows worldwide, Git AI capitalizes on this familiarity to lower the barrier for adopting AI technologies. It provides open APIs, ensuring customers maintain sovereignty over their code, models, and data, a critical factor for those cautious about vendor lock-in and data portability.

Benefits of Git AI include:

• Accessibility on various platforms, including our partners' GPU Clouds and public clouds such as Lambda Labs and Phoenix Systems, as well as support for private Kubernetes environments.
• A rich starting library of models, including open-source Large Language Models (LLMs), transformers, and image synthesizers, facilitating a broad range of AI applications.
• The ability to maintain AI sovereignty, enabling users to freely move their experiments from cloud vendors to on-premises setups thanks to Git AI's open APIs.

Licenses

The Git AI library is available for free under the permissible Elastic License v2. You can use, extend, or deploy models from the library as long as you do not provide Git AI as a managed service to another party.

Product

Git AI is available with our GPU cloud partners and through our website hosted on public cloud. Please give us feedback at support@codedepot.ai.

List the cloud services:

• Git extension for AI - users can create and manage Git AI repositories in CLI using a Git-native API.
• Git AI library - enables users to instrument their experiments.
• Depot CLI - a command line interface where users can create and manage training experiments as well as manage their clusters.
• Web app - enables users to run jobs and track experiments directly through the user interface, while also providing tools for comparing models and experiments.
• GraphQL API - enables users to query in real-time their data available through the web app and the command line tools.