Developers, agents, and startups are spinning up entire products mere days after ideation.

But the tools teams use to understand what they've shipped haven't changed enough to keep up. Someone still has to manually tag features. Someone maintains the dashboards. Someone remembers to check them. 

Now, it’s too much to keep up with. Some teams fall behind on instrumentation and quietly drift because their data no longer reflects what shipped. Others abandon analytics altogether, deciding they'll circle back when things slow down. Either way, you end up refactoring features and shipping new workflows based on a version of your product that no longer exists.

Over the last few months, we've gone back to first principles and asked: what should product analytics and adoption tools actually look like in this new world of working? That answer led us to Novus, the product agent for AI-first teams. 

Novus is now available, for free

Novus is the product agent for teams shipping software with AI. It connects to your codebase, understands product structure and user behavior, continuously monitors changes, and turns observations into actionable recommendations or automations. Today, it's in open beta and free to get started.

It’s designed for the way product engineers, technical PMs, and fast-moving teams are building today. After six weeks of design partner feedback, we’ve brought even more powerful features to the open beta. 

Here’s what you can do with Novus: 

  • Sets up in minutes. Novus scans your codebase, maps the product, and submits a pull request to install instrumentation, with design partners seeing setup times under 10 minutes.
  • Builds living product memory. Once Novus understands your codebase, it creates Memory: a dynamic map of your product that captures the features you’ve built, the personas using them, and the workflows those personas are trying to complete.
  • Connects directly to coding agents. Novus integrates with tools like Claude Code and Cursor via MCP, so your coding agent already understands what you’re building, where it lives, and who it’s for.
  • Finds issues before you do. Novus continuously monitors your product, combines real usage data with business context, and flags problems before they become obvious. It surfaces a narrative explanation, root cause, and recommended code-level fix plan.
  • Turns insight into action. You can hand the fix plan to Claude yourself or let Novus create the pull request for you. Either way, Novus proposes changes and you approve them before anything merges.
  • Reviews UX before shipping. When a pull request comes in, Novus evaluates the diff through a user-experience lens using behavioral data. It can catch issues like broken flows, missing redirects, or dead ends before they reach production.
  • Automates launches. When a new feature lands in a PR, Novus can automatically set up rollout plans, segmented feature flags, in-product guides, and goal tracking. You review and approve the plan, and Novus handles the execution.
  • Keeps product context current. As your product changes, Novus stays in sync with your code and usage patterns, so the system’s understanding of your product remains accurate over time.

Novus handles human-generated flows and agentic ones. If your product includes AI agents, Novus instruments those too: capturing full conversations, tool calls, and whether users actually completed their goals. This is what’s next.

Building in public with powerful partners like Anthropic

Novus works alongside Claude Code, and we're partnering with Anthropic to deepen that integration. 

When Memory connects to your coding agent, it gives the model a live picture of how your software works and how people use it. The difference between an agent that builds things and an agent that builds the right things.

Try Novus.ai for free today, and tell us what you’re doing with it.