Getting Started with TokenBoard

This guide walks through the normal TokenBoard flow: install the local collector, connect your browser account, let usage sync, and read the leaderboard.

The product is intentionally simple. You do not need to paste prompts, upload project folders, or configure a cloud agent. TokenBoard reads supported local AI coding telemetry, extracts usage metrics, and sends only the numbers needed to build your profile.

1. Install the collector

From the TokenBoard homepage, copy the install command and run it in your terminal.

curl -fsSL https://tokenboard.dev/install.sh | sh

The installer sets up a local TokenBoard collector. On supported systems, it also configures a user-level background service so usage can continue syncing after restarts.

If your site uses a different production domain, use the install command shown on the site itself. The command on the homepage is the source of truth.

2. Connect your account

After installation, the collector opens a browser-based device login flow. Sign in with your TokenBoard account and confirm the device code.

This step links the local machine to your web account. It does not grant TokenBoard access to your repositories. It only allows the collector to report usage metrics under your profile.

3. Let recent activity sync first

TokenBoard prioritizes freshness. Recent activity should appear before the collector finishes deeper historical backfill.

That means your profile can become useful quickly:

• New sessions show up first.
• Older local sessions may appear later.
• Totals can increase as backfill catches up.
• The device page shows connected machines and privacy controls.

If the leaderboard looks quiet immediately after install, give the collector time to find supported local telemetry.

4. Read the leaderboard

The leaderboard has several useful views:

• Users shows who is reporting the most AI coding usage.
• Models shows which models account for usage volume.
• Skills shows observed work categories when the collector can infer them safely.

Sort by tokens, sessions, messages, or estimated cost depending on the question you are asking. Tokens show volume. Sessions show cadence. Cost gives a rough business-friendly translation. Model mix shows how experimental or concentrated the usage is.

5. Open your profile

Your profile is where the leaderboard becomes personal. It includes high-level totals, model mix, activity rhythm, and daily details.

Use it to answer practical questions:

• Which days did I actually use AI coding tools?
• Which models dominate my workflow?
• Did my usage pattern change after a new project or workflow?
• Is my activity steady enough to share as proof?

The profile is designed to be shareable without showing the private conversations behind the numbers.

6. Manage devices and privacy

Visit the devices page when you want to inspect or revoke connected collectors. This is also where TokenBoard keeps device-level controls, such as whether a machine is allowed to upload future usage.

Revoke a device if you no longer use that machine, if you are rotating access, or if you want to stop reporting from a particular environment.

Troubleshooting

If usage does not appear:

1. Confirm you signed in and completed the device confirmation page.
2. Check that your AI coding agent is supported by the current collector.
3. Leave the collector running long enough for the first sync cycle.
4. Open the devices page and confirm the machine is still connected.
5. Reinstall from the homepage command if the local service was removed.

What not to expect

TokenBoard does not read or display your private prompts. It does not show assistant responses. It does not publish command output or source code. It does not guarantee billing-grade accounting.

It gives you a clear usage profile: enough to compare, share, and understand your AI coding rhythm without turning private work into public content.