Carbon OffsetNow rolling out

Account for the carbon behind your AI.

FormWise estimates the carbon footprint of your AI usage and retires verified carbon credits to match it, with proof you can show your clients.

From a model call to a retired credit.

Four steps, all of them visible in your dashboard. Nothing is retired on a promise: credits are retired against usage that has been estimated, billed, and paid for.

  1. 01

    Estimate

    Every model call your agents and flows make gets a CO₂e estimate, using per-model emission factors from the open-source EcoLogits methodology.

  2. 02

    Bill

    After each month closes, the CO₂e you accrued is billed on your next invoice at a flat per-tonne rate, shown in your dashboard before you opt in.

  3. 03

    Retire

    Once that invoice is paid, verified carbon credits covering it are retired, in a batch with other organizations.

  4. 04

    Prove

    Retirement proof is recorded on-chain and listed in your retirement history, with a transaction reference you can open and share.

You stay in control of the spend.

Carbon Offset is a setting, not a surcharge. You decide whether it runs, how much it can cost, what it covers, and who pays.

Opt in when you are ready

Credit retirement is off until you turn it on. Turn it off before a month closes and that month is skipped entirely: nothing billed, nothing retired.

Set a monthly spend cap

Pick a cap or leave it open. When the cap is reached, retirement pauses for the rest of the month. Your usage is never blocked.

Choose what is covered

Every flow and agent is included by default. Exclude any one of them from its own editor.

Pass the cost through

Plans you sell can pass the carbon offset cost on to subscribers, so each customer covers the footprint of their own runs.

Straight answers.

The method is open. Read the EcoLogits methodology.

How is the footprint calculated?

We apply per-model emission factors derived from the open-source EcoLogits methodology, an ISO 14044 life-cycle assessment of AI inference, using the midpoint of its published uncertainty range. Newly released models it does not cover yet are estimated conservatively by model class.

Are these measured emissions?

No. The figures are estimates, not measurements. Nobody outside a model provider can meter the energy behind a single call, so we use a published, open method and say plainly that it is an estimate.

When are credits actually retired?

After the month closes and the invoice that includes your carbon charge is paid. Retirements run in a shared batch, and your history shows your portion alongside the on-chain proof once it is confirmed.

Does this make my AI carbon neutral?

We do not make that claim, and we suggest you do not either. What you can say, accurately, is that the estimated footprint of your AI usage is matched by retired, verified carbon credits, and that you can show the proof.

Put your AI to work, and show the receipts.

7-day free trial. Onboarding call included. Turn Carbon Offset on from your dashboard whenever you are ready.

Start free trial