Welcome to the Oobeya ecosystem, Claude Code.
Oobeya now includes a native Claude Code integration for usage monitoring, token cost tracking, and engineering impact analysis. It expands our AI assistant connections alongside GitHub Copilot, Cursor, and OpenAI Codex.
The connection brings Claude Code telemetry into AI Impact, including its Metrics and Cost & Credits tabs. Engineering leaders can review assistant activity and spending alongside code quality, review delays, rework, and delivery performance.
This integration ships with FE 2.0.645 and BE 2.0.994. See the 2 November release notes for the full update, including the redesigned Team Scorecard and new data-health symptoms.
Claude Code usage monitoring across plans
The integration uses OpenTelemetry to collect Claude Code activity. Teams can connect supported Claude Code clients across plans, rather than depending on an Enterprise-only usage export. Anthropic documents the underlying collection mechanism in its Claude Code monitoring guide.
In AI Impact → Metrics, select Claude Code in the Type filter and choose a reporting period. The activity summary brings the main usage signals together:
- Participation: active users and sessions.
- Requests: user prompts, model requests, and request errors.
- Time: developer time and assistant time, shown separately.
- Activity: available commit and pull request counts, with trends over time.
The detailed views extend that picture with model requests and input, output, cache-read, and cache-creation tokens, plus tool runs and success rates. These measures help teams investigate changes in consumption or execution patterns without treating every request as equivalent work.
Claude Code activity summary and usage trends in AI Impact → Metrics.
For example, rising request errors can explain why an assistant is taking more attempts to complete work. A change in the balance of input and cached tokens can prompt a closer look at how teams use context. The reporting gives platform teams a place to investigate those patterns before expanding a rollout.
Claude Code cost monitoring, with billing context
Open AI Impact → Cost & Credits to review Claude Code spending for the selected month. The view brings together gross usage value, daily consumption, month-end usage forecasts, total users, and cost per active user.
There are two financial measures to keep distinct:
| Measure | What it tells you |
|---|---|
| List-price usage estimate | The value of Claude Code consumption calculated using list prices. |
| Enterprise billed amount | When a Claude Enterprise API key is configured on the data source, the billed amount supplies net cost; the difference is shown as a discount. |
A usage estimate is useful for comparing consumption. It should not automatically be read as an additional charge on a subscription invoice. Anthropic's cost management documentation explains the difference between estimated token costs and billing.
Claude Code usage and financial reporting in AI Impact → Cost & Credits.
For an engineering leader, Claude Code token cost tracking becomes more useful when the discussion moves beyond the bill: is consumption concentrated in a few workflows? Does a spending increase coincide with faster reviews, or with more rework? Which teams need a closer look at usage before their allocation grows?
Explore AI Usage & Cost Monitoring for the broader financial view across supported assistants.
Connect Claude Code activity to engineering outcomes
Assistant usage describes what happened inside the tool. AI Impact brings that activity into the engineering workflow.
Review coding efficiency, PR time to merge, bug counts, lead time, and quality indicators alongside AI-related trends. Outcome Signals summarizes how many tracked engineering outcomes improved, stayed stable, or worsened compared with the previous period.
Engineering metrics and Outcome Signals provide context for changes in assistant activity.
A shorter merge time is useful context; a simultaneous rise in bugs deserves investigation. Teams can follow those signals into Team Performance, DORA Metrics, and quality analysis to understand the work behind the trend.
These comparisons show relationships over time. They help teams investigate AI's contribution without presenting a change in usage as proof of a productivity gain. Learn more about this reporting in AI Impact.
Connect Claude Code to Oobeya
Start with the Claude Code integration in Oobeya:
- Open Integrations → Claude Code → New Data Source. Select OpenTelemetry, name the connection, and configure its Ingest token.
- Configure Claude Code's OpenTelemetry export using the connection settings for your Oobeya environment. Follow the Oobeya integration documentation and Anthropic's OpenTelemetry setup guide.
- If you need Enterprise billed-cost reporting, configure the Claude Enterprise API key in the data source's Anthropic API key field.
- Once telemetry has been collected, open AI Impact, select Claude Code, and review Metrics and Cost & Credits for the same team scope and period.
The telemetry connection supplies activity and consumption data. The optional Enterprise key adds financial context; it is not a prerequisite for collecting OpenTelemetry activity across plans.
Already using Claude Code with Blamely?
Blamely's Claude Code support already captures line-level AI/human code attribution in supported workflows. That Git-based provenance feeds Oobeya's AI Usage analysis.
The new native connection adds a different source of context:
| Connection | What it contributes | Where to review it |
|---|---|---|
| Native Claude Code integration | Assistant activity, token consumption, and cost reporting | AI Impact, Metrics, and Cost & Credits |
| Existing Blamely attribution | Line-level AI/human origin and Git provenance | AI Usage and code-attribution analysis |
Use the two together to connect what the assistant consumed, where AI-generated code entered the codebase, and what happened to engineering quality and delivery.
One ecosystem for your AI coding tools
Claude Code joins the native reporting ecosystem alongside GitHub Copilot, Cursor, and Codex. Oobeya supports cost monitoring for GitHub Copilot, Claude Code, and Codex, while retaining each provider's reporting definitions.
The result is a shared place to review AI coding activity, spending, and engineering outcomes across the tools your teams choose.
Ready to connect Claude Code? Explore the integration, read the release notes, or schedule a walkthrough with your team's usage and cost questions.
Continue Exploring
Written by Sukru Cakmak
Sukru Cakmak is the Co-Founder & CTO of Oobeya. He works closely on the platform's technical direction, engineering intelligence capabilities, and the practical challenges of measuring software delivery, developer productivity, and AI-assisted development across modern SDLC environments.
