On February 20, 2025, GitLab made headlines with the launch of version 17.9, a major update designed to significantly enhance user experience and functionality across its platform.

Among the standout features is the introduction of GitLab Duo Self-Hosted, allowing organizations to deploy large language models (LLMs) entirely within their own infrastructures!

This groundbreaking feature enables companies to utilize GitLab Duo for code suggestions and chat directly from models hosted on-premises or in secure private clouds.

The update supports cutting-edge open-source Mistral models via vLLM or AWS Bedrock, Claude 3.5 Sonnet on AWS Bedrock, and even models from OpenAI on Azure OpenAI—all while maintaining top-tier security protocols.

Organizations eager to harness LLM technology without the fear of exposing proprietary code or sensitive data will find this rolling out as a game-changer, particularly in regulated industries like finance, healthcare, and government.

It paves the way for AI adoption while ensuring complete control over critical data—an essential method for tackling strict compliance and regulatory challenges.

Joel Krooswyk, GitLab’s Federal CTO, shared insights with DevOps.com, revealing that, despite the rising popularity of the software-as-a-service (SaaS) edition, many firms are still inclined towards self-hosting to meet specific regulatory demands.

He emphasized the crucial role that self-hosted solutions play in helping DevOps teams address data privacy issues effectively.

Additionally, executing LLMs on-premises or within a private cloud allows organizations to sidestep latency caused by external API calls, particularly beneficial for real-time AI applications.

By processing AI tasks internally, compliance to regulations such as GDPR, HIPAA, and SOC 2 becomes more manageable, ensuring that sensitive information stays securely within the company's control.

GitLab 17.9 is not just about AI; it also introduces powerful capabilities for secure coding practices integrated directly within the development pipeline.

LLMs will now assist teams in identifying security vulnerabilities and suggesting best practices while even autonomously rectifying issues before code merges—ushering in a new era of 'shift-left security' where safety and compliance are embedded earlier in the development lifecycle.

And that's not all! This latest release brings remarkable enhancements, allowing users to run multiple GitLab Pages sites with parallel deployments—making simultaneous updates a breeze and drastically improving overall efficiency!

In terms of collaboration, the integration capabilities have been boosted significantly, enabling developers to add project files directly into Duo Chat while utilizing popular integrated development environments (IDEs) like Visual Studio Code and JetBrains.

This level of integration is set to expand teamwork and productivity like never before.

Moreover, GitLab 17.9 automates the cleanup of older pipelines, optimizing project maintenance by managing storage efficiently.

This means users can wave goodbye to cluttered repositories filled with outdated information, ensuring that resources are utilized more effectively.

As AI technology continues to evolve, GitLab is stepping up to assertively facilitate its adoption within the development community.

This new rollout signals a substantial leap in how teams can proficiently harness LLMs to enhance their efficiency and productivity in unprecedented ways.

Stay tuned for more exciting developments!