Introducing Claude Code Analytics in Kindor

Engineering teams are adopting AI coding tools quickly. The harder question is no longer whether a tool is available - it is how it is being used across the organization, what it costs, and how that usage connects to the work teams deliver.

Today, we are introducing Claude Code Analytics in Kindor.

The new integration brings Claude Code usage into the same view as the contributors, teams, commits, pull requests, and delivery metrics you already use in Kindor. It gives engineering leaders a privacy-conscious way to understand adoption and usage patterns without adding a Kindor agent to developer machines.

From AI usage to an operational view

Claude Code Analytics helps answer practical questions such as:

• How broadly is Claude Code being adopted across teams?

• Which models are being used, and where is spend concentrated?

• How much AI-assisted code is being created or modified?

• What share of commits can be attributed to Claude Code sessions?

• Which teams or contributors may benefit from enablement and shared practices?

The goal is not to reduce engineering work to a single productivity number. AI usage, code volume, and commits are useful signals, but they need context. By connecting these signals to Kindor's existing contributor and delivery model, teams can examine AI adoption alongside their real operational data.

A team-level view of adoption, cost, and delivery

The Claude Code Usage dashboard provides an organization and team-level view of adoption over time. Leaders can see active contributors, daily active users, total cost, cost per merged pull request or merge request, AI-attributed commits, model mix, programming languages, agents, and skills used.

It also groups contributors into relative usage cohorts, non-users, light, medium, and heavy users, so teams can explore adoption patterns alongside average commits and merged pull requests. These cohorts are intended to prompt useful questions and learning, not to rank individuals.

Go beyond the usual metrics and see which skills and agents are being used across your teams as well as which is the most widely used language when working with agents.

A contributor-level view for context

We included an additional dashboard at the contributor level that allows you to understand in a deeper way how Claude is being used, to easily detect patterns and the impact AI assistants are having in the day to day work of individual contributors.

Built on OpenTelemetry, with a zero-footprint rollout

The integration uses Claude Code's OpenTelemetry support to send selected telemetry to Kindor. Configuration is distributed centrally through Claude.ai managed settings, so developers do not need to install a Kindor agent or configure their machines individually.

Claude.ai Teams or Enterprise and the Owner or Primary Owner role are required to enable this managed configuration.

AI usage data needs careful boundaries. Kindor is designed to collect the operational signals needed for adoption analysis while excluding the content of developers' work.

Kindor does not collect prompts, source code, raw API bodies, or full shell command lines. It captures selected signals such as sessions, token usage, cost, code-edit decisions, AI-authored lines of code, and commits or pull requests created in Claude Code sessions.

Get Started

Claude Code Analytics is now available for Kindor customers. To enable it, contact the Kindor team to receive your organization-specific token and integration guide.