GitLab launches dedicated AI gateway for regulated enterprise workflows
- GitLab Dedicated AI Gateway enables single-tenant agentic AI for regulated firms
- Secrets Manager extends credential control to Kubernetes and Terraform tools
- Bulk SAST remediation allows triage of thousands of vulnerabilities in one action
- Usage caps for GitLab Credits provide monthly spending ceilings for AI features
- Flow Creator Agent generates runnable flows from plain language descriptions

*this image is generated using AI for illustrative purposes only.
GitLab Inc. has released its Dedicated AI Gateway, enabling regulated enterprises to run agentic AI workloads within single-tenant environments. The update addresses data residency concerns by keeping AI-processed data inside existing security boundaries.
Dedicated AI Gateway
The AI Gateway for GitLab Duo Agent Platform is now generally available. It allows customers running sensitive software delivery workloads on GitLab Dedicated to execute the agent platform within their specific region. Users can connect their own models for inference without moving data outside their trusted infrastructure.
This capability targets audited environments where standard multi-tenant AI services may not meet compliance requirements. The gateway ensures agentic workloads follow the same isolation model as the rest of the software development lifecycle.
Secrets Manager Expansion
GitLab Secrets Manager is now in limited availability as a paid add-on billed through GitLab Credits. It securely stores credentials used both inside CI pipelines and by external infrastructure tools like Kubernetes, Terraform, and OpenTofu.
Every secret is scoped to the environment, branch, and protection status of the job requiring it. This unified permission model eliminates the need for separate credential management systems outside the platform.
Security Remediation Automation
Bulk SAST False Positive Detection and Agentic SAST Vulnerability Resolution are now in beta. These features allow security teams to triage thousands of open vulnerabilities in a single action.
The system provides a confidence score for each finding. Confirmed risks receive ready-to-merge fixes, reducing manual coding effort. The tool continues to automatically triage new critical and high-severity findings as they arrive.
Flow Creator Agent
Flow Creator Agent, now generally available, converts plain language descriptions into complete, runnable custom flows. This removes the need for users to learn the Flow Registry schema manually.
Flows generated via Agentic Chat run under scoped service accounts with composite identity. Enabling these flows requires the Maintainer role or higher, ensuring administrative control over automation deployment.
Additional GitLab 19.3 Features
GitLab Credits usage caps are now generally available. Administrators can set monthly ceilings on agentic AI spend at the subscription level or per user via the GraphQL API. This prevents unexpected overages in AI consumption costs.
Restricted visibility for custom agents and flows per GitLab group is also generally available. This makes assets accessible to members across multiple projects within that group, supplementing existing per-project and public visibility options.
What the Numbers Show
The simultaneous release of usage caps alongside new agentic AI features indicates a strategic focus on cost predictability. By allowing granular control over GitLab Credits spend, GitLab addresses a primary barrier to enterprise adoption: unbounded AI inference costs. This suggests the company is prioritizing financial governance tools to accelerate uptake among budget-conscious regulated sectors.
How might GitLab's single-tenant AI Gateway impact the competitive landscape against multi-tenant cloud providers like AWS or Azure in highly regulated industries?
What are the potential long-term cost implications for enterprises as GitLab Credits usage caps become standard, and will this model encourage or discourage heavy agentic AI adoption?
Could the expansion of GitLab Secrets Manager into a paid add-on fragment the DevSecOps toolchain, or will it successfully consolidate credential management workflows?

































