Claude Code Usage Dashboard: Free & Paid Trackers

The right Claude Code usage tracker depends on the question you need to answer: how much of your allowance remains, which project consumes the most tokens, or what your engineering team costs each month. This guide compares free and paid options and explains how to connect usage to developers and projects.

Keep usage limits, estimated token costs, and actual charges separate. A dollar estimate for subscription-covered usage is not automatically an extra charge. Anthropic explains the distinction in its cost guide.

Which Claude Code usage dashboard should you use?

Free and paid Claude Code trackers and monitoring options
Tool	Cost model	What it tracks	Best fit
Claude Code /usage	Included with existing access	Session and plan usage	Individual developers checking consumption
ccusage	Free, open source	Local history, tokens, project breakdowns and estimated costs	Comparing projects on your machine
Claude Code Usage Monitor	Free, open source	Live consumption, forecasts and warnings	Monitoring active sessions
Anthropic organization analytics	Eligible paid plans and Console access	Organization activity and individual usage	Team adoption reporting
Enterprise project tags	Eligible Enterprise plans; beta	Supported spend by project tag or cost code	Project spend allocation
SigNoz	Self-hosted or paid cloud	Centralized telemetry and custom dashboards	Teams configuring shared monitoring

Free refers to the tracking software. Claude usage, hosting, and implementation may still carry costs. SigNoz publishes its cloud pricing separately.

How do I check my Claude Code usage?

Run this command inside Claude Code:

/usage

It provides session information and, where applicable, subscription usage details. Use it for an immediate consumption check. Confirm actual spend in your billing account, especially when your organization mixes subscriptions, usage credits, and API billing. Claude Code usage documentation.

How do I track Claude Code usage by project for free?

Start with ccusage. After installation, group local daily usage by project, filter a project, or export the report:

ccusage daily --instances
ccusage daily --project my-project
ccusage daily --instances --json

Reports include tokens and estimated costs, with model breakdowns and date filters available. Current versions support multiple coding tools; select the Claude Code source for a Claude-only report. See the ccusage reporting guide.

For a team-wide report, collect exports from the relevant devices, associate them with developer identities, and map local folder names to consistent project names. One developer’s backend folder and another’s payments-api folder may belong to the same repository. Missing local history means missing usage.

How do I see Claude Code usage by developer?

Start with Anthropic’s native analytics if your organization has access. Eligible administrators can open Analytics → Claude Code; Console users have a separate Claude Code page. Reporting includes organization activity and individual usage, with permissions depending on the account and role. Native analytics access and metrics.

For custom developer, team, and project breakdowns, aggregate OpenTelemetry data in a shared monitoring platform. Configure consistent identities and attribution labels before building reports. Some authentication methods require you to supply user identity explicitly. Claude Code telemetry reference.

SigNoz provides a Claude Code monitoring setup. For Ganivra setup, follow our guide to tracking Claude Code cost per developer and our managed settings rollout guide.

Can I allocate Claude Code spend to a project or cost code?

Eligible Enterprise organizations can use project tags, currently in beta. For supported Claude Code sessions, merge the attribution field into the repository’s existing .claude/settings.json:

{ "attributionProject": "PROJECT_ID" }

Administrators can then review Organization settings → Usage → Spend by project. Check plan, deployment, and version requirements first: API-key deployments and several other routes are excluded, seat-covered usage is not counted, and earlier spend is not retroactively tagged. Project tag setup and limitations.

Define whether a project means a repository, client engagement, or finance cost code. Track unassigned usage so missing labels do not silently distort the totals.

How can I monitor Claude Code usage in real time?

Claude Code Usage Monitor provides live terminal monitoring, forecasts, and warnings. With uv installed, its documented setup is:

uv tool install claude-monitor
claude-monitor

Use forecasts to guide your workflow and Claude’s own usage information to check account limits. For centralized monitoring, freshness depends on telemetry export and ingestion intervals.

What should a Claude Code cost dashboard show?

Developer and team: who generated the usage.
Project: the repository, client, or initiative that owns it.
Model and token type: input, output, cache writes, and cache reads.
Time period: comparable daily, weekly, and monthly trends.
Cost basis: estimated token cost, allocated subscription cost, or actual billed charges.

Adding estimated costs to subscription fees can double-count the same activity. Review them separately. Higher token use can reflect larger tasks or more automation; token totals alone do not measure developer productivity.

A usage monitor answers what was consumed. A budgeting workflow connects consumption to ownership and charges. Explore AI coding cost tracking with Ganivra to track coding-tool usage across developers and projects, and use the team chargeback guide when allocating costs.

Also using Codex? Compare OpenAI’s Codex usage dashboard and what OpenTelemetry adds.

For production APIs, gateways, customer margins, and billing, see AI cost tracking tools compared.

Claude Code usage tracking: questions and answers

Select a question to read the answer.

How do I check my Claude Code usage?

Run /usage inside Claude Code to check session information and available plan usage details. The session dollar estimate is not necessarily your subscription bill. Confirm actual charges in your billing account.

Is there a free Claude Code usage tracker?

Yes. ccusage is a free, open-source tool for local usage reports, including project breakdowns and estimated costs. Claude Code Usage Monitor is another free option for live terminal monitoring and consumption forecasts. Claude usage itself may still cost money.

How can I track Claude Code usage by project?

With ccusage installed, run ccusage daily --instances to group local usage by project. Select the Claude Code source if your reports include other tools. For team reporting, normalize project names across devices. Eligible Enterprise organizations can also use project tags for supported spend.

How can I see Claude Code usage by developer?

Use Anthropic's organization analytics with an eligible account and role, or aggregate Claude Code OpenTelemetry data in a monitoring platform. Centralized reporting needs consistent developer identities; local reports alone do not automatically include every developer or device.

Does Claude Code token cost match my bill?

Not necessarily. Token-cost estimates can value usage at API rates even when a subscription covers it. Actual charges depend on your plan, usage credits, provider, and contracted rates. Keep estimates separate from subscriptions and billed charges to avoid double-counting.

What is the best Claude Code usage dashboard for a team?

Start with native analytics for organization adoption reporting. Choose centralized telemetry when you need custom developer, team, or project breakdowns. Compare identity mapping, project attribution, retention, exports, and the distinction between estimated costs and billed charges before choosing a platform.

Can I monitor Claude Code usage in real time?

Claude Code Usage Monitor provides live local monitoring and consumption forecasts. Centralized OpenTelemetry dashboards update according to export and ingestion intervals. Forecasts are estimates; use Claude's account usage information to check plan limits.