Terms

You need to be able to enter a binding agreement, and you are responsible for what happens under your account — including runs started by anyone you invite to your team.

Sign-in is by emailed link or Google. Keep control of that mailbox; it is the key to the account and to everything the agents can reach.

People are on a plan; teams are billed for usage. A personal account chooses a plan and is charged monthly in advance for the credits it grants. A team has no per-person fee, no minimum and no subscription floor: adding or removing people changes nothing on its invoice, and a month in which no agent ran is a month with nothing to pay.

The free plan takes a card, and starts the paid plan by itself. It grants a fixed number of credits with no time limit. There is nothing to cancel on a schedule and no charge while those credits last — but spending them starts the paid plan automatically, charging the card on file at the price shown when you chose it. Cancelling before the credits run out costs nothing.

Usage is what your agents run.Agents call models, and that is metered. For a team the amount for the period is shown on the team page as it accrues, broken down per person, and billed monthly in arrears for the month just finished. For a personal account it draws down the plan’s credits, which roll over.

Ceilings are yours to set. Every run has a hard limit, and admins can set monthly caps per person and for the team. We are not responsible for spend within limits you configured, and a run refused for hitting one is the system working.

Fees are non-refundable except where the law requires otherwise, or where we have plainly billed you for something that did not run. Cancel any time; access continues to the end of the paid period.

Do not use traven to break the law, to access systems you are not authorised to access, to send unsolicited bulk messages, or to produce material that harms people — malware, harassment, deception designed to defraud.

Connecting a credential asserts that you are allowed to. Agents act with the permissions you grant them, and granting access to a system you do not control is a problem you have created rather than one we can detect.

What agents produce is yours. We claim no ownership of your briefs, files or results, and do not use your content to train models.

It is not guaranteed to be correct. Language models produce confident, plausible, wrong output. Anything consequential — code that ships, numbers in a deck, a message sent to a customer — needs a human to check it. That is not a disclaimer bolted on; it is why the product shows you every step and every file rather than handing over an answer.

You are responsible for what you do with the output, including anything an agent sends to a connected system on your instruction.

The service is provided as it is. We do not warrant that it will be uninterrupted, that a run will succeed, or that a model provider we depend on will stay available.

To the extent the law allows, our total liability for any claim is capped at what you paid us in the three months before it arose. We are not liable for indirect or consequential loss — lost profit, lost data, business interruption.

We may change these terms. Material changes get notice by email to account owners before they take effect, and continuing to use the service after that is acceptance.

We can suspend an account for non-payment or for the use described above. Where it is safe and lawful to do so, we will say why first.

admin@traven.ai. See also the privacy policy and the security page.

Written plainly rather than by a lawyer, and not reviewed by one. If you need a negotiated agreement, a DPA or specific terms, write and we will deal with it properly instead of pointing at this page.