Tools Connect: Anthropic (Beta)
Allstacks now supports a Beta integration with Anthropic, these metrics allow teams to monitor performance, usage, and cost of Anthropic services.
Important Notes
Beta Status: This integration is in Beta and hidden behind a feature flag. Please contact your Allstacks Customer Success Manager (CSM) representative to enable it.
Active User: Data presented in Allstacks will be from the last 90 days.
Live Data: Metrics refresh every time you reload the page, as fresh data is fetched on each load.
Note: rate limiting can occur upon refresh. Rate limit is based on Tier level on the Anthropic account: https://docs.claude.com/en/api/rate-limits#rate-limits
Versioning: If you will be changing versions, it will require you to reconfigure the connection with a new Admin Key & Version value.
Prerequisites
You will need to:
Need an Anthropic Console Account
Create an Admin Key
Anthropic Version will be needed for the API to pull in the data.
Version can be found here: https://docs.claude.com/en/api/versioning
How We Connect
Once you have the feature flagged enabled for the Anthropic Beta by your Customer Success Manager you should see
Anthropic Usage Metricsbecome available in yourConfiguresection to navigate to.
Once on the
Anthropic Usage Metricsdashboard page, top-right, click theConfigurebutton. You will receive a pop-up window which will allow you to enter in your Admin API key and Anthropic Version, save it to start pulling in data.
We use the following endpoint to pull data which you can also validate to ensure Allstacks is representing the same data you are seeing in Anthropic.
Usage API https://docs.claude.com/en/api/admin-api/claude-code/get-claude-code-usage-report
Cost API https://docs.claude.com/en/api/admin-api/usage-cost/get-cost-report
Metrics We Show from Anthropic
Once configuration is completed, Allstacks displays the following metrics sourced from Anthropic:
1. Total Active Users
Number of users active in the past 90 days.
Why it matters: Helps measure adoption and ongoing engagement across the organization.
2. Total Lines of Code Accepted
Total lines of code added and removed in the last 90 days.
Why it matters: Indicates overall coding activity and developer productivity trends.
3. Total Spend
Estimated Total USD spent in the past 90 days.
Why it matters: Tracks cost efficiency and helps identify opportunities to optimize usage.
4. Cost Over Time
Daily cost trends across services and models.
Why it matters: Reveals usage spikes, spending patterns, and cost-saving opportunities.
Note: Values under $0.01 display as zero.
5. Lines of Code Over Time
Daily lines of code added and removed.
Why it matters: Shows momentum in development and how AI assistance impacts coding speed.
6. User Activity Over Time
Daily active users and usage patterns.
Why it matters: Highlights engagement trends and helps spot drops in adoption early.
7. Model Usage Distribution
Bar chart of total cost by AI model, with dynamic colors per model type.
Why it matters: Identifies which AI models drive the most cost and usage for better resource planning.
8. Token Usage Over Time (by Model)
Token usage trends across Claude models, grouped by model type.
Why it matters: Tracks model efficiency and helps align token use with cost and performance goals.
9. Tool Actions Distribution
Daily accepted and rejected AI tool actions.
Why it matters: Shows how often AI suggestions are accepted, helping evaluate model accuracy and value.
10. User Usage Breakdown
Detailed usage and cost stats for each active user.
Why it matters: Provides visibility into top contributors, power users, and potential training needs.