Needs AI usage in the Scope 3 report with a method an auditor will accept.
What they get: One report per provider and per project, a cited methodology with a stated uncertainty band, and a signed receipt and audit pack for each closed month.
Measure what you use
Every token your teams spend, on every provider, reported as carbon you can defend.
Connect a provider, install the CLI or add the extension, and the dashboard turns the usage into one report per model, per provider and per project. Paid plans close each month with a signed receipt and an audit pack.
Live · 13 Sep, 17:19 UTC
Tokens · 365 days
27.4B
365 days to 13 Sep 2026
CO₂e · 365 days
1.7 t
lifecycle · methodology v1.2
Providers
9
connected
Usage by provider
- Anthropic57.5%
- OpenAI25.7%
- Vercel AI Gateway9.0%
- Z.ai4.5%
Usage over time
Oct 2025 to Aug 2026
Receipt · signedFetched 13 Sep 2026, 17:19 UTC
Explore the live dashboard
8
providers and gateways you can connect from the dashboard
203,000
verified credits issued
Crbon Labs, 2026
4,500+
SDK and MCP downloads with no paid marketing
Crbon Labs, 2026
28
percent uncertainty band on every figure, stated
What is the Dashboard?
The CrbonFree dashboard is the web application at dashboard.crbonfree.com where AI usage from provider connections, including the OpenAI usage API, coding agents and the browser extension becomes one report. It shows tokens, cost, energy and CO2e per model, per provider and per project, calculated with a published methodology, and on paid plans it closes each month with signed credit retirement receipts and an audit pack.
Last checked against the product on 5 September 2026. Carbon figures follow methodology v1.2.
At a glance
- Runs
- In the browser, at dashboard.crbonfree.com
- Sources
- OpenAI, Anthropic, OpenRouter, AWS Bedrock, Vercel AI Gateway, Cloudflare AI Gateway, Google Vertex AI, Microsoft Azure Foundry, the crbon CLI
- Reports
- Tokens, cost, energy, CO2e and water, per model, provider, project and agent
- Ranges
- 7 days to a custom window; CSV and JSON export
- Methodology
- v1.2, three layers, uncertainty stated at about 28%
- Proof
- Signed receipts with serial numbers, monthly audit packs with a manifest hash
- Plans
- Free $0 audit-only; Starter $29; Growth $99; Scale $299
- Signs in with
- The browser, or a device code for headless machines
One report across providers, projects and coding agents
The dashboard shows total and daily tokens, compute cost, energy, CO2e and water for the range you pick, from the last seven days to a custom window, filtered by project. My coding agents adds the usage the CLI syncs from developer machines, per agent, per model and per week, next to the health of each connection. Any view exports as CSV or JSON.
Per model and per provider, with the methodology behind every figure
Usage by model sits on the dashboard and inside every project, and the coding-agents view lists the provider beside each model. The carbon card cites the methodology the figure was calculated with, every calculation is pinned to the factor version it ran under, currently v1.2, and each billing period records the versions used to close it.
Receipts and audit packs an auditor can check
When a period closes on a paid plan, the credits retired against it are recorded on a signed receipt with the credit serial numbers, a PDF and a public verification page. Audit packs bundle a month of signed receipts into a zip with a manifest and its hash, and trust assets add a badge and a certificate for the claims the receipts support.
Built for a whole organisation
An organisation holds members, roles and invitations. Projects group connections and get their own thirty-day report. Any member can create their own CLI key from My coding agents, account API keys for the SDKs and the MCP server live under settings, and a device code signs in a headless machine.
How the Dashboard moves your data
- 1
Usage arrives
A provider key or a cloud sign-in on the Connections page, the CLI syncing coding agents from developer machines, or the browser extension reporting chat sessions.
- 2
The backend attributes and computes
Each event is matched to a model tier and converted to energy and CO2e through three inspectable layers under methodology v1.2, with upper and lower bounds.
- 3
The dashboard reports
Tokens, cost, energy, CO2e and water appear per model, per provider, per project and per agent, for the range you choose, ready to export.
- 4
Receipts and audit packs close the period
On a paid plan the period closes, credits are retired, and a signed receipt with serial numbers and a monthly audit pack record the result under the factor version in force.
How to get started
- 1
Create an account
Sign up at dashboard.crbonfree.com and create or join an organisation. If you want to look first, the sample account opens the same views on live public usage with no login.
- 2
Connect a source
On the Connections page, paste a usage key from a provider, sign in to a cloud account, or install the crbon CLI on the machines that run coding agents. Organisation admins manage provider connections; every member can connect their own CLI.
- 3
Read the first report
The dashboard fills in as usage arrives. Provider usage APIs report with a delay of one to two days, so the first complete day lands shortly after you connect. Pick a range, filter by project, open the model breakdown and export it if someone else needs the numbers.
This is where everything lands
Every other CrbonFree surface either feeds the dashboard or reads from it. The CLI and the browser extension put usage in. The TypeScript and Python SDKs, the MCP server and the REST API read the same usage, emissions, receipts and audit packs back out, so a figure quoted anywhere matches the one on this page.
See all the surfacesWhat it does not do
The carbon figures are model-tier estimates with a stated uncertainty band of about 28%, not meter readings from the provider. Retiring credits never changes the measured footprint; the two are kept apart on purpose. Signed receipts, audit packs and trust assets are part of the paid plans, and the free plan is audit-only. Notifications are marked coming soon in settings, and the browser extension card on the Connections page still says coming soon even though the extension is in the Chrome Web Store. Provider usage APIs lag by one to two days, and late telemetry is reconciled before a period closes.
Who reaches for the Dashboard
Wants to see which models, providers and coding agents are behind the token bill without building an internal tool.
What they get: Usage by model and by provider on one screen, the coding agents per developer and per week, and CSV or JSON export for anything that needs to go further.
Has to reconcile AI spend, quotas and the cost of retiring credits every month.
What they get: Billing periods with tokens, carbon and cost, overage charges against the plan quota, invoices and the receipts that show what each retirement covered.
Which CrbonFree surface do I need?
Three of them put usage in: the dashboard through provider connections, the CLI for coding agents and the browser extension for chat sessions. The other four read the numbers back into your own code. The Dashboard is highlighted.
| Surface | Meters usage | Reads numbers back | Runs | Needs | Language |
|---|---|---|---|---|---|
| Dashboardthis page | In the browser | An account; the sample account opens without one | None | ||
| CLI | status | On your machine | An account; device code for headless machines | Node 20 or newer | |
| Browser extension | estimates | In Chrome | Optional, per the store listing | None | |
| TypeScript SDK | Node 18 or newer | An API key | TypeScript or JavaScript | ||
| Python SDK | Python 3.8 or newer | An API key | Python | ||
| MCP server | Beside your MCP client | An API key from login | Any MCP client | ||
| REST API | Over HTTPS | An API key | Any |
Questions
Questions about the Dashboard.
Short answers. The package page and the API reference have the long ones.
Open the sample accountOpenAI shows it on its own platform usage page, per project and per model, with a delay of a day or two. CrbonFree reads the same usage API with an admin key you create and shows the same tokens beside Anthropic, OpenRouter, Bedrock, Vertex AI and Azure Foundry, with cost, energy and CO2e per model, so one report covers every provider.
Yes. The sample account at crbonfree.com/demo opens the dashboard, the coding-agents view and the connections page on live public usage from OpenRouter, with no login. Cost, energy, carbon and water in it are modeled estimates, the same way they would be for your own usage.
For each closed period, a signed receipt with the CO2e retired, the credit serial numbers, a PDF and a public verification page served independently of the dashboard. A monthly audit pack bundles the signed receipts into a zip with a manifest and its hash. Every figure is pinned to the factor version it was calculated under, and the methodology page documents the math, the sources and the uncertainty.
Eight today: OpenAI, Anthropic, OpenRouter, AWS Bedrock, Vercel AI Gateway, Cloudflare AI Gateway, Google Vertex AI and Microsoft Azure Foundry. Cohere, xAI, Groq, Together AI, Replicate and Ollama are listed as coming soon. Coding agents on developer machines report through the crbon CLI rather than a provider key.
The free plan is audit-only: usage tracking and carbon estimates with no recurring charge. Starter is $29 a month for up to 100 million tokens, Growth is $99 for up to 1 billion and Scale is $299 for up to 3 billion, and each paid plan adds signed receipts, monthly audit packs and trust assets, with overage billed per million tokens above the quota. Larger quotas are handled on an enterprise plan.
Any member of the organisation can create their own CLI key from My coding agents, so a team can start metering without waiting for an admin. Account API keys for the SDKs and the MCP server are created under settings and shown once. Provider connections and projects are managed by organisation admins, who also invite members and assign roles.
Goes well with
All productsCLIMeter your local AI coding agents automatically, synced to the dashboard every hour.Browser extensionEstimates token usage in your AI chat sessions across ChatGPT, Gemini and the Claude apps, including Claude Code. Guest mode keeps everything on your machine, and signing in keeps the history.TypeScript SDKA typed client for your usage, emissions, receipts and audit packs from Node.js or TypeScript.
See your own number instead of ours.
However your team already runs AI, that is how you connect. Paste a read-only provider key, drop in an SDK, add the MCP server, install the CLI or switch on the browser extension, and the first measured tokens reach your dashboard within seconds.
Create your free account
Free plan · first tokens in seconds