GitLab Duo Agent Platform Is Now Generally Available: What It Means for Your Teams

As of January 15, 2026, GitLab Duo Agent Platform is generally available (GA). The release marks a shift in how AI is used across the software development life cycle (SDLC): from individual coding assistance to AI agents that operate across planning, security, delivery, and operations—inside a single DevSecOps platform.

According to the GitLab Global DevSecOps Report 2026, organizations face an “AI paradox”: while AI speeds up coding, fragmented tools and manual handoffs still cost teams an average of seven hours per person per week.

GitLab Duo Agent Platform addresses this gap by embedding AI agents across the SDLC and reducing rigid stage boundaries. Instead of managing handoffs step by step, teams coordinate multiple agents that work in parallel—shortening feedback loops and reducing delivery delays.

As a GitLab Partner, we’re happy to walk you through the core capabilities of the Duo Agent Platform, its pricing particularities, and what they mean in practice.

Features of GitLab Duo Agent Platform

GitLab Duo Agent Platform combines multiple capabilities designed to work together rather than as standalone AI tools. Agents run asynchronously and in parallel, allowing planning, development, review, and pipeline fixes to happen without strict sequencing.

Specialized AI Agents

At GA, the platform includes three core agents, with more planned and support for customer-built agents:

• Planner Agent (GA)
  Breaks epics into structured tasks and suggests priorities based on project context. Use cases include backlog refinement and sprint planning.
  
• Security Analyst Agent (GA)
  Reviews vulnerabilities, explains risk in plain language, and recommends remediation order. Useful for security reviews and compliance preparation.
• Data Analyst Agent (Beta)
  Interprets project metrics and highlights trends for retrospectives and release insights.

Flows

Flows turn multi-stage work into repeatable processes. Available GA flows include:

• Issue to Merge Request
• Convert to GitLab CI/CD
• Fix CI/CD Pipeline
• Code review
• Software Development guidance within GitLab

Custom, organization-defined flows are available in beta, allowing teams to apply internal standards consistently.

MCP: Connecting Your Toolchain

 The MCP Client is now GA, enabling Duo Agent Platform in IDEs to securely connect to external systems like Jira, Slack, and Confluence.

• Available at GA: Secure connection to external MCP-compatible systems (Jira, Slack, Grafana, etc.) with group-level access controls.
• Currently in Beta: The GitLab MCP server capability, which will allow for even deeper integration, is in beta with more features planned for future releases.

AI Catalog and Agentic Chat

The AI Catalog serves as a central place to manage agents and flows—publishing approved configurations, assigning them to projects, and connecting custom agents to internal systems.

Agentic Chat provides multi-step reasoning using full GitLab context—issues, merge requests, pipelines, and security findings—directly in the GitLab UI. It supports code generation, refactoring, testing, documentation, and project summaries without switching tools.

Pricing Model: Usage-Based, Not Seat-Based

Duo Agent Platform introduces GitLab Credits as a shared usage currency.

Here’s How Credits Work

• 1 GitLab Credit = $1 list price
• Credits apply across usage-based GitLab AI products
• No contract amendment is required
• Volume discounts are available through pre-commit paths

How Customers Pay

1. Included Credits for GitLab existing users:
   
   Premium: 12 credits per user per month
   Ultimate: 24 credits per user per month
   
   Note: These included credits are available for a limited time as part of a launch promotion and are subject to change.
2. On-Demand Credits: Usage beyond the included amount is billed at $1 per credit monthly.
3. Pre-Commit Pool: Organizations can pre-purchase a shared pool of credits with an annual commitment and built-in volume discounts.

This approach avoids the problem of paying for unused seats while still supporting power users and team-wide adoption.

What Existing Duo Pro and Duo Enterprise Customers Should Know

Customers with active Duo Pro or Duo Enterprise subscriptions retain their existing entitlements. Any remaining per-seat licensing value can be converted into GitLab Credits for use with the Duo Agent Platform, protecting their current investment. All feature sets included in Duo Pro and Duo Enterprise remain available with unlimited usage, as they are not subject to usage billing.

In practice, teams can move from a single AI assistant to agent-based workflows across the SDLC without losing previously purchased value.

Self-Hosted Models: Where Things Stand

GitLab has outlined an industry-first option designed for customers operating in regulated or tightly controlled environments:

• A seat-based SKU for self-hosted models alongside Duo Agent Platform
• Support for GitLab Self-Managed deployments, including on-premises, private cloud, and air-gapped environments
• Intended for organizations requiring private AI usage

There is no public timeline yet for an offline pass, but this direction confirms GitLab’s commitment to AI adoption without forcing customers into public cloud-only models.

Cloudfresh’s Role as a GitLab Partner

Moving to an agent-based platform introduces new variables—from credit consumption to custom flows. Cloudfresh acts as your technical bridge, ensuring your transition to the Duo Agent Platform is aligned with your team’s real needs.

As a GitLab Select & Professional Services partner, our team supports you at every stage of adopting Duo Agent Platform:

• Advising on pricing models and credit estimation
• Helping teams transition from Duo Pro or Enterprise
• Designing custom agents and flows aligned with internal processes
• Supporting governance, usage tracking, and cost visibility
• Keeping you in the loop on Beta versions, GA updates, and roadmap changes

Let’s connect to evaluate how GitLab Duo Agent Platform fits your delivery, security, and governance priorities—and to plan adoption with clarity from day one.