Building the Future of Telemetry in the Open

In the modern world of agentic AI, innovation doesn't wait for the next major product release cycle. With every iteration, new tools such as AI coding assistants, mobile frameworks, and novel backend architectures are continuously rewriting the rules of productivity for developers and engineering teams, making it increasingly difficult to fully embrace all the latest offerings as well as understand how they are actually impacting your efficiency and effectiveness. In addition, the rapid adoption of these new tools often creates a gap within enterprise observability platforms, which are the very systems required to keep your technologies stable, secure, and cost-effective. The reality is that it often takes time for formal integrations to become available which are able to monitor the quality and effectiveness of these critical emerging tools. So, what happens in the early development period between a disruptive technology's debut and its formal enterprise rollout? Developers are forced to fly blind, rely on patchwork workarounds, or build their own bespoke tools from scratch. But there is a better way.

Welcome to New Relic Experimental, our open-source incubator designed to bridge the gap and accelerate enterprise observability. This is where New Relic engineers and global community members collaborate in real-time to build, test, and share the bleeding-edge telemetry solutions your modern stack demands today.

What is New Relic Experimental?

Think of it as a collaborative sandbox where New Relic employees (Relics) and passionate engineers from across the globe build, test, and share cutting-edge observability solutions. A completely open-source repository where  projects don’t carry strict service-level agreements, but a space for community and collaboration where practitioners can  find  something with real value and  immediate utility. This is where we share our rapid innovation to solve new challenges in real-time, working openly with the community to figure out the best ways to add observability to  tomorrow's tech stack.

Whether you are looking to monitor a niche Node.js framework or trying to wrangle the costs of your team's AI coding tools, there’s a good chance someone in the experimental community is already building a solution.

Spotlight: What We Are Building Right Now

The repository is constantly growing, but here are several examples of standout experimental projects currently in active development that you can clone and experiment with:

• (Coming soon!) New Relic AI Coding Observability. Engineering teams are adopting AI coding assistants like GitHub Copilot, Cursor, and Claude Code at unprecedented speeds. However, this rapid adoption has created a massive blind spot. FinOps teams are dealing with opaque monthly invoices, and engineering leaders are struggling to prove whether these tools actually improve developer velocity or just introduce new anti-patterns like prompt thrashing. The AI Coding Observability project extends your existing New Relic platform directly into the developer workflow. By installing lightweight hooks, it captures every AI tool call, model interaction, and cost event. It provides the vendor-neutral telemetry you need to govern multi-tool AI spend, objectively prove developer ROI, and capture compliance-grade audit logs of what your AI coding assistants are really doing.
• Machine Learning Performance Monitoring. Deploying a machine learning model is only half the battle, ensuring it stays accurate in the wild is the real challenge. ML models can degrade silently over time as real-world data drifts away from training data. The Machine Learning Performance Monitoring project provides the vital instrumentation needed to track model drift, prediction accuracy, and inference latency in production. It allows data science and engineering teams to ensure their ML investments continue to deliver reliable, performant business outcomes long after deployment.
• New Relic React Native Module. Mobile observability requires deep, framework-specific hooks that generic web monitors simply can't provide. As React Native continues to dominate cross-platform mobile development, teams need better ways to see inside their applications. The React Native Module is an experimental integration engineered to bring robust, native-level crash reporting, distributed tracing, and mobile performance telemetry directly into your React Native applications, helping you eliminate blind spots between your mobile frontends and backend services.
• NextJS Integration. Modern Node.js architectures demand specialized, low-friction instrumentation. NextJS has become a go-to framework for building scalable server-side applications, but manually instrumenting its complex dependency injection and decorators can be tedious. The NextJS Integration project provides drop-in observability for the NestJS framework. It automatically surfaces the performance of your controllers, services, and middleware, allowing you to trace requests through your NextJS architecture without writing heavy, custom instrumentation code.

Help Us Shape the Future of Observability

We open-source these experimental projects because we firmly believe the best tools are built in the open. The New Relic Experimental repository isn't just a place to download code, it's a place to collaborate. Highly adopted and community validated experimental projects frequently graduate to fully supported New Relic Community or Public Catalog assets. By jumping in early, you have the opportunity to directly influence the roadmap of the tools you rely on.

Ready to experiment? Head over to New Relic Experimental. Fork a repository, deploy an experimental hook to see what it can do for your stack, or open a pull request. Let’s build the next generation of observability together.