Novus Review (2026): Pricing, Features & Honest Verdict

TLDR

Novus by Pendo is a product agent that connects to your codebase, auto-instruments your app, and continuously watches real usage. When it spots broken flows or dead ends, it rates the severity and opens a pull request with a fix. It launched this week on Product Hunt in free open beta. Best for: product engineers and technical PMs shipping fast. Price: free beta, no paid pricing yet. Rating: 6.6/10.

What is Novus?

Novus is a product agent built to catch usability problems before your users hit them. You connect it to your codebase, and it scans the repo, maps your product, and submits a pull request to install instrumentation, with design partners reporting setup under 10 minutes. From there it builds what Pendo calls Memory: a living map of the features you’ve shipped, the personas using them, and the workflows those people are trying to complete. It then monitors live usage continuously, flags issues like broken flows, missing redirects, and dead ends, assigns each a severity, lays out an investigation plan, and opens a GitHub pull request with a proposed fix you review.

It launched this week on Product Hunt in early June 2026, landing at #8 product of the day with 142 votes, and it is in open beta and free to get started. The notable part is the pedigree: Novus comes from Pendo, an established product analytics company, and was built in partnership with Anthropic. It also plugs into Cursor and Claude Code via MCP, so your coding agent already understands what you are building and who it is for. That positioning, an analytics veteran shipping an AI agent that writes the instrumentation and the fixes for you, is what makes Novus more interesting than a typical week-one launch.

What Are Novus’s Key Features?

Codebase connection and auto-instrumentation

The headline feature is that you don’t wire up analytics by hand. Novus connects to your codebase, scans it, and opens a pull request to install instrumentation automatically. Design partners report setup times under 10 minutes, which, if it holds up, is a dramatic improvement over the days of manual event tagging that product analytics usually demands. It collects the signals teams often forget to track, closing the gap between the events you meant to capture and the ones you actually did.

Continuous usability monitoring

Once instrumented, Novus monitors your product daily. It combines real usage data with the business context in its Memory map to flag problems proactively: funnel drop-offs, broken flows, missing redirects, and dead ends. Each flagged issue gets a severity rating so you know what to fix first. The pitch is to surface issues before they become obvious in your metrics, rather than waiting for a dashboard review or a support ticket spike.

Investigation plans and pull requests

When Novus catches something, it cross-references your GitHub repo, configuration, and live usage to investigate what changed, then proposes a fix directly as a pull request. Crucially, nothing merges without you. You review the investigation findings and approve the PR before anything ships. It can also catch issues like broken flows and dead ends in a PR before they reach production, acting as a usability check in your pipeline rather than only a post-launch monitor.

Rollout automation and AI agent integration

Beyond fixing problems, Novus can set up the scaffolding around new features. When a feature lands in a PR, it can automatically create rollout plans, segmented feature flags, in-product guides, and goal tracking, leaning on Pendo’s analytics heritage. It integrates with Claude Code and Cursor through MCP, so the coding agents you already use share context with Novus. This is where the Pendo and Anthropic partnership shows, tying instrumentation, monitoring, and AI-assisted coding into one loop.

How Much Does Novus Cost?

Novus is free to get started right now, because it is in open beta. There are no published paid tiers, and Pendo has not announced pricing for general availability. So the honest answer is that today it costs nothing, but that is a beta offer rather than a permanent free plan, and you should expect paid plans once the product matures.

That makes a clean price comparison impossible, but it is worth knowing the neighborhood. PostHog, the open-source product analytics and session replay tool, has a free tier and then usage-based pricing. LogRocket, focused on session replay and frontend monitoring, starts around $69/mo. Pendo’s own established analytics suite uses custom enterprise pricing that typically runs into four or five figures a year. Novus is a different shape of product, an agent rather than a dashboard, so when it does price, where it lands relative to those tools will say a lot about who it is really for.

The practical read: free beta is a low-risk way to evaluate the concept, but do not build long-term budget assumptions on it. If Novus proves its detection and prices like Pendo’s core suite, it becomes an enterprise purchase. If it prices like PostHog, it stays accessible to smaller teams. Until Pendo announces, treat the cost as an open question and the free access as a trial window.

Who is Novus Best For?

Use Novus if you ship product quickly and want usability problems caught automatically instead of discovered weeks later in a dashboard. Product engineers and technical PMs who are tired of manually instrumenting events will get the most from the auto-instrumentation and PR-based fixes. Teams already using Cursor or Claude Code via MCP fit especially well, since the agents share context out of the box.

• You want analytics instrumentation set up automatically, not hand-tagged over days.
• You ship often and need to catch broken flows and dead ends before users do.
• You are comfortable letting an agent open PRs that you review before merging.

Skip Novus if your organization can’t connect a production codebase and GitHub to a third-party agent, which is a hard no for many security and compliance teams. Skip it if you need a mature, proven analytics platform with years of references today, because Novus is days old and its detection accuracy is unproven. And since pricing is unknown, hold off if you need budget certainty before you commit.

Best Novus Alternatives

PostHog

PostHog is the open-source leader for product analytics, session replay, and feature flags, with a generous free tier and usage-based pricing after that. It gives you deep, self-hostable analytics and a huge feature set, but you still instrument and analyze largely by hand. Novus’s bet is that an agent doing the instrumentation and proposing fixes saves more time. Pick PostHog for control, breadth, and a proven track record, Novus for automation.

LogRocket

LogRocket focuses on session replay and frontend monitoring, starting around $69/mo, and is excellent for watching exactly what a user did before something broke. It surfaces frontend errors and friction with mature, battle-tested tooling. It does not auto-instrument from your codebase or open fix PRs the way Novus aims to, so it is the safer, more established choice if you mainly want replay and error visibility rather than an agentic fix loop.

Pendo (core suite)

Pendo’s own flagship analytics and in-product guidance suite is the enterprise-grade sibling to Novus, with custom pricing and a long history serving product teams. It covers adoption analytics, in-app guides, and feedback at scale. If you want proven, comprehensive product analytics today rather than a brand-new agent, the core Pendo platform delivers that, while Novus is the experimental, code-connected, AI-driven layer on top.

Final Verdict: Is Novus Worth It?

Novus is one of the more credible week-one launches we’ve looked at. The idea, an agent that instruments your product from the codebase, watches real usage, flags usability issues with severity, and opens PRs to fix them, is genuinely useful, and the under-10-minute setup would remove the single most tedious part of product analytics. Coming from Pendo and built with Anthropic, it has more pedigree than the average new tool, and the MCP integration with Cursor and Claude Code is a smart way to slot into how teams already build.

The caveats keep the rating grounded. It launched this week in open beta, so detection accuracy and false-positive rates are unknown, and an agent that flags the wrong things or opens noisy PRs would cost more time than it saves. Pricing is unpublished, so the free access is a trial window, not a long-term plan. And connecting a production codebase to a third-party agent is a real trust and security decision. We’d try the free beta on a non-critical product, judge the quality of its flagged issues and proposed fixes, and watch for pricing news before relying on it. Promising and worth testing, but unproven, so go in as an early adopter, not a committed buyer.