TokenUse is a token and cost observability platform for developers and teams using AI coding agents. It tracks supported sessions, breaking down token usage and costs by provider, model, project, and session in real time.

How does TokenUse track token usage?

TokenUse uses a lightweight CLI tracker that runs in the background on your machine. It monitors supported local event sources, normalizes the data, and streams it to your dashboard via the API. Setup takes under 2 minutes.

Which AI models does TokenUse support?

TokenUse supports Anthropic Claude Code and OpenAI Codex session tracking. Cursor Agent device detection is available today; Cursor session tracking and cost attribution are on our roadmap.

How does pricing work?

TokenUse offers four plans: Hobby (free forever), Team ($29/month for higher monthly usage), Scale ($99/month for higher-volume usage), and Enterprise (custom pricing). All plans include real-time tracking and the CLI.

How do I install TokenUse?

Install via Homebrew (brew install tokenuse/tap/tokenuse), npm (npx tokenuse), or curl. Then run 'tokenuse login' to authenticate and 'tokenuse tracker run' to start tracking. The CLI runs as a background daemon on macOS and Linux.

TokenUse encrypts data in transit (TLS 1.3); at-rest encryption of stored data is on our roadmap and not yet in place. You maintain full control over your data and can export or delete it at any time.

What data does TokenUse collect?

TokenUse collects token counts (input, output, cache read, cache write), model identifiers, project and file paths, and session metadata. By default it also captures the text of the prompts you run to power per-prompt analytics—this is on by default and you can turn it off anytime (Settings → Data, or `tokenuse config set prompts.enabled=false`). Detected secrets are scrubbed before upload, prompts are kept per your retention setting, and we never collect model responses or your file contents.

Can I use TokenUse with my team?

Collaboration features are on our roadmap. Today, each account tracks one user's local agent usage.

TokenUse – Token & Cost Observability for Developers

TokenUse tracks AI token usage and cost in real time so teams can control spend and move faster.

TokenUse provides per-model, per-project, and per-session visibility with alerts, reporting, and budget controls for developers and teams using AI coding agents.

Quick Answer

TokenUse tracks AI token usage and cost in real time so teams can control spend and move faster. This page explains practical next steps for tokenuse – token & cost observability for developers.

What You Can Do With TokenUse

Understand tokenuse – token & cost observability for developers in the context of AI token and cost operations.
Connect product capabilities to measurable outcomes for engineering and finance.
Use linked pages below to continue setup, evaluation, or procurement workflows.

Implementation Checklist

Confirm the owner and measurable outcome for this workflow.
Verify instrumentation captures model, project, and environment dimensions.
Define alert and escalation rules before broad rollout.
Link implementation docs and pricing assumptions for team review.

How It Works

Install the TokenUse CLI with one command.
Authenticate your account to connect tracking.
Start the tracker to capture AI token usage automatically.
View dashboards to analyze spend by model, project, and session.

Who It's For

Engineering Leaders — Track AI usage across teams and prioritize optimization work.
Platform Teams — Instrument AI operations with consistent telemetry and budget controls.
Finance & FinOps — Forecast AI spend, allocate costs, and maintain budget discipline.

Implementation Context

Operational Snapshot

TokenUse helps teams connect AI usage with ownership, budget controls, and operational accountability.

TokenUse tracks AI token usage and cost in real time so teams can control spend and move faster. Use this page to align engineering, platform, and finance workflows around shared spend visibility.

Common Workflows

Instrument AI requests with lightweight setup so usage data is available quickly.
Group usage by model, owner, and project to surface spend drivers.
Apply alerts and reporting to keep cost behavior predictable.
Export data and share reports across engineering and finance teams.

Common Questions

What is TokenUse?: TokenUse is a token and cost observability platform for developers and teams using AI coding agents. It tracks supported sessions, breaking down token usage and costs by provider, model, project, and session in real time.
How does TokenUse track token usage?: TokenUse uses a lightweight CLI tracker that runs in the background on your machine. It monitors supported local event sources, normalizes the data, and streams it to your dashboard via the API. Setup takes under 2 minutes.
Which AI models does TokenUse support?: TokenUse supports Anthropic Claude Code and OpenAI Codex session tracking. Cursor Agent device detection is available today; Cursor session tracking and cost attribution are on our roadmap.
How does pricing work?: TokenUse offers four plans: Hobby (free forever), Team ($29/month for higher monthly usage), Scale ($99/month for higher-volume usage), and Enterprise (custom pricing). All plans include real-time tracking and the CLI.
How do I install TokenUse?: Install via Homebrew (brew install tokenuse/tap/tokenuse), npm (npx tokenuse), or curl. Then run 'tokenuse login' to authenticate and 'tokenuse tracker run' to start tracking. The CLI runs as a background daemon on macOS and Linux.
Is my data secure?: TokenUse encrypts data in transit (TLS 1.3); at-rest encryption of stored data is on our roadmap and not yet in place. You maintain full control over your data and can export or delete it at any time.
What data does TokenUse collect?: TokenUse collects token counts (input, output, cache read, cache write), model identifiers, project and file paths, and session metadata. By default it also captures the text of the prompts you run to power per-prompt analytics—this is on by default and you can turn it off anytime (Settings → Data, or `tokenuse config set prompts.enabled=false`). Detected secrets are scrubbed before upload, prompts are kept per your retention setting, and we never collect model responses or your file contents.
Can I use TokenUse with my team?: Collaboration features are on our roadmap. Today, each account tracks one user's local agent usage.

Platform: Tracking, attribution, alerts, and reporting workflows.
Pricing: Plan limits, billing options, and upgrade paths.
Documentation: Quickstart, CLI reference, API integration, and troubleshooting.
Trust and Security: Security, privacy, and data handling practices.