New in NeoLoad: Delivering Core Web Vitals in 2025.3 and expanding AI-driven MCP capabilities

In what has already been an innovation rich year, Tricentis NeoLoad continues to evolve to meet modern performance engineering needs. The 2025.3 release brings user‑centric insights with Core Web Vitals in RealBrowser which can leverage greater scalability through the recent availability of our high‑capacity cloud load generator. We are also continuing to build on our recently updated NeoLoad Web frontend with the addition of new and enhanced features.

Our AI capabilities have certainly not slowed either since the release of our NeoLoad Model Context Protocol (MCP) earlier this year. The team has been hard at work delivering a wide variety of new MCP actions to extend AI‑driven automation across infrastructure management, analysis workflows, and more.

There’s a lot to cover, so let’s start with NeoLoad 2025.3 …

RealBrowser + Core Web Vitals with NeoLoad 2025.3

NeoLoad 9 introduced RealBrowser, our solution for browser‑based testing. By using real headless browsers to replicate user journeys, RealBrowser closes the gap left by protocol‑level tests by validating rendering, script execution, third‑party tags, and dynamic content. Since launch, we have continued to enhance RealBrowser, and with NeoLoad 2025.3 it now captures Core Web Vitals, providing user‑centric metrics directly within performance test results.

Core Web Vitals are Google‑defined metrics that quantify real‑world user experience across three dimensions: Largest Contentful Paint (LCP) measures loading performance, Interaction to Next Paint (INP) measures responsiveness, and Cumulative Layout Shift (CLS) measures visual stability.

Poor scores in any of these areas translate into frustrating digital experiences, lower engagement, and potential SEO impact. By capturing LCP, INP, and CLS under load, NeoLoad makes it possible to evaluate both system performance and user experience quality in the same test, reducing tool sprawl and enabling teams to resolve issues before they reach production.

Key benefits include:

• Capture Google‑defined UX metrics during load tests for speed, responsiveness, and stability.
• Identify user experience issues before release that affect conversion, trust, and engagement.
• Validate SEO outcomes by testing against Core Web Vitals thresholds as part of routine performance runs.
• Reduce reliance on separate UX tools and simplify reporting across teams.

High‑capacity cloud load generator

Core Web Vitals and browser‑based testing are critical for capturing true end user experience of your customer facing applications, but browser-based testing can be more resource intensive than typical back-end protocol tests. To address this, we recently made available a new option to configure a high-capacity cloud load generator, which you can learn more about here.

Coupling RealBrowser tests with the high-capacity load generator allows you to:

• Run large browser‑based tests in the cloud with 16 CPU and 64 GB RAM instances.
• Confidently simulate peak and burst conditions for complex user journeys.
• Simplify operations by scaling on demand instead of managing hardware.

New features in the enhanced NeoLoad Web

With NeoLoad 2025.1, we also made available our new and enhanced interface for NeoLoad Web, supported on the backend by our updated fully open V4 API architecture. Our aim was not only to deliver a user experience and aesthetic to be more consistent with other Tricentis solutions, but to maintain the capabilities users have come to expect from NeoLoad Web while delivering new innovation as well.

One new example is the availability of an analysis treeview. From the Results>Values tab, users can aggregate and sort through various metrics such as requests and transactions, as well as a shared containers view and the ability to trace timeseries instantly, allowing for the generation of graphs on the fly. For those that work in NeoLoad GUI, this has been a requested view that we carry over into NeoLoad Web.

There is always more coming in our continuous delivery model here, so stay tuned!

Model Context Protocol in NeoLoad

Earlier this year, we were the first to market with a set of MCPs across our products, including NeoLoad. As a quick refresher, the Model Context Protocol provides a secure, standardized way for users and AI assistants to interact with NeoLoad through natural language. The initial rollout allowed users to run and monitor tests, query results, and generate reports conversationally, while also enabling integration across other Tricentis MCP‑enabled tools and third‑party systems.

Taking advantage of the fully open API architecture we released in NeoLoad 2025.1, we are rapidly expanding MCP actions available to our customers across key performance testing workflows. Recently, we have added a number of capabilities across infrastructure and zone management, deeper analysis and reporting capabilities, and more.

MCP infrastructure actions

The new MCP infrastructure tools remove much of the complexity from configuring and managing load testing infrastructure. Instead of navigating manual setup steps or scripting, teams can use natural language prompts to create, size, and control environments on demand.

This makes it faster and simpler to align infrastructure dynamically with changing test requirements, ensuring that performance testing can scale and adapt without delays or specialized expertise. The result is more efficient environment management, more reliable test execution, and less time lost to setup hurdles.

Some advantages for your workflows include the ability to:

• Manage environments through natural language prompts that reduce setup time.
• Create, size, and control zones and providers on demand.
• Enable or disable cloud zones and verify provider connectivity for reliable execution.

MCP analysis actions coupled with Augmented Analysis

We recently published a blog covering this which you can check out here, but to recap, we have also expanded MCP capabilities for results analysis by enabling teams to retrieve detailed elements, drill into transactions and requests, and explore logs and events with conversational prompts. Interval‑based analysis highlights how performance changes over time, while Augmented Analysis automatically surfaces anomalies and likely root causes. Together, these capabilities speed triage, improve communication, and make insights easier to consume across roles.

This will help your teams to:

• Retrieve detailed results, transactions, and request values quickly.
• Run interval‑based analysis to spot degradations or spikes across phases.
• Leverage Augmented Analysis to surface anomalies and root causes automatically.
• Generate stakeholder‑ready reports instantly to accelerate decisions.

Conclusion

While NeoLoad 2025.3 brings valuable new metrics to further increase capabilities in RealBrowser, our latest MCP actions continue to expand the world in which a substantial amount of performance testing workflows can be driven and managed through natural language from an LLM.

Looking ahead, the new MCP capabilities mark just the beginning. They are opening the door to more intelligent automation in performance testing, and they lay the foundation for our upcoming Performance Agent that will take on even more of the setup, execution, and analysis work. With these advancements, teams can expect faster feedback, less manual effort, and performance testing that adapts in step with the pace of modern delivery.

In the meantime, join the NeoLoad 2025.3 and expanded MCP release webinar to learn more about the innovation we’ve discussed in this blog and see these capabilities in action.