Today, we’re excited to announce the IBM DevOps Loop 2025.12 (2.0) release—our next major step in helping teams move from AI-assisted development to fully agentic, automated software engineering workflows. This release expands the power of Loop as an open, AI-driven platform designed for modern, fast-moving enterprises that need both agility and governance.
DevOps Loop continues to evolve into the foundation for augmented, AI-first delivery, helping teams orchestrate work, automate developer tasks, and elevate productivity across the entire software lifecycle.
What’s New in DevOps Loop 2025.12
The 2025.12 (2.0) release introduces enhancements across AI integration, orchestration, developer experience, and platform extensibility—all built to support MCP-driven agentic workflows.
Expanded AI & Agentic AI Capabilities
This release significantly broadens the AI ecosystem available within DevOps Loop.
- Integration with Claude Desktop and VS Code
- Support for Gemini and Claude model frameworks
- Conversation support added to Loop Genie, enabling more natural and contextual orchestration
- Enhanced MCP server and local MCP tools, improving accuracy, responsiveness, and local code manipulation
- Added support for Ollama as an AI provider, giving teams more flexibility in model deployment and cost management
These updates empower development teams to leverage the AI tools they prefer while maintaining enterprise-level governance, cost visibility, and security.
Modernized User Experience & New Workflow Enhancements
We’ve refreshed the UI to make Loop cleaner, faster, and more intuitive—and we’ve added functionality that gives teams more control over how projects evolve.
Add New Resources Directly into Existing Loops
Teams can now seamlessly expand active Loops with:
- Control repository integration with build pipelines
- Test applications for coverage and QA workflows
- Additional asset types for more complete orchestration
Quick Search: Expanded and More Intelligent
Quick Search now indexes additional data sources to speed up exploration and analysis:
- Build documents
- Pull request documents
- Commit documents
- Deployment documents
This gives teams richer visibility across work items, code artifacts, and delivery pipelines.
Why This Release Matters: Preparing for Agentic Engineering
Across the industry, a clear shift is underway: teams want the speed of individual AI-assisted development, without sacrificing the auditability and control required by the enterprise.
Many organizations today still operate across a wide productivity spectrum—some using waterfall, others fully agile/DevOps, some exploring AI generation, and a growing number transitioning into agentic workflows.
Four core challenges continue to push teams toward MCP-driven automation:
- Toolchain and data sprawl
- Slow, costly testing cycles
- Manual handoffs that break visibility
- AI sprawl—multiple models with no unified strategy
DevOps Loop 2025.12 directly addresses these challenges by:
- Unifying tools, AI providers, automation, and project data
- Providing a single MCP-powered orchestration layer
- Supporting third-party orchestrators like Watson Orchestrate, GitHub Copilot, Claude, and others
- Enabling agents to plan, analyze, generate, and execute work
- Offering built-in governance and reporting
This is the bridge between AI-assisted coding and true agentic engineering—where AI agents understand context, make decisions, and execute workflows across the entire lifecycle.
A Platform Built for Openness and Real-World Delivery
DevOps Loop continues to be an open platform:
- Integrates with Jira, Aha!, GitHub, VS Code, and other industry tools
- Connects with IBM technologies such as ELM, ClearCase, WebSphere, Instana, HashiCorp, and more
- Automates planning, creation, testing, deployment, and measurement
- Provides built-in test automation, data virtualization, and release orchestration
- Supports any software target: cloud, mainframe, mobile, APIs, AI services, and distributed systems
With 2025.12, teams can now pair all this automation with the intelligence of MCP-based agents—unlocking measurable productivity gains across development, QA, and operations.
A Glimpse of What’s Next
Even today, MCP-powered agents within DevOps Loop can:
- Generate and refine code
- Analyze builds and test failures
- Plan sprints
- Update work items
- Manipulate local code
- Trigger deployments
- Manage extensions and resources
- Communicate directly with Loop Genie or third-party orchestrators
This release sets the stage for what’s coming in 2026: deeply interconnected agentic workflows that run with minimal human orchestration and maximum governance.
Get Started with IBM DevOps Loop 2025.12
The 2025.12 release is available now!
For documentation, tutorials, or upgrade guidance, visit the IBM DevOps Loop product page or contact your IBM representative.