Terms of Service

The service

Aevrin scans MCP (Model Context Protocol) servers for security issues using established open-source tools, and reports the results to you. It does not modify, install, or execute the servers it scans in any way that affects your systems beyond the scan itself.

Aevrin also reads the configuration of coding agents on a machine you run it on and reports what each agent has been allowed to do, and it operates Aevrin MCP, a hosted MCP endpoint your own agent reaches with your Aevrin API key. Through Aevrin MCP your agent can search the capabilities an Aevrin administrator has reviewed and published, and for each one either Aevrin runs the capability on its servers and returns the output, or Aevrin returns the instructions that administrator wrote and your agent carries them out itself.

Where your agent carries out instructions on your own machine, that work runs with your privileges and under your control, not inside Aevrin's sandbox, and you remain responsible for it. Where Aevrin runs a capability, it runs in an isolated sandbox that can reach only the hosts named in that capability's approved allow-list. The output of a capability is produced by third-party code: it is data to read, not instructions to act on, and Aevrin labels it as such.

Accounts and usage limits

Aevrin has two plans, Free and Pro. Each includes a fixed monthly allowance of credits, reset on a rolling monthly cycle from your signup date, and each metered action spends a fixed number of credits: an MCP scan costs 50 credits, an AI code review 50 (available on Pro only), an Aevrin MCP capability run 10, an agent posture upload 5, and a capability read 1. We may adjust the allowances and costs from time to time; published numbers at the time you use the service apply.

The Free plan includes 1,500 credits per month, one capability run at a time, and one monitored device. It does not include AI code review or report export.

The Pro plan includes 50,000 credits per month, AI code review, two capability runs at a time, up to ten monitored devices, and report export. A one-time top-up purchase adds 5,000 credits that do not expire and are spent after the monthly allowance. A capability read is one retrieval of the details and instructions of a capability; searching for a capability is free and is not counted. Where a capability allows a recent result to be reused instead of being run again, that answer costs the same as a run and spends none of your monthly compute time.

Running a capability on Aevrin's servers is not switched on yet. Until it is, such a request returns an error and no capability run is counted against either plan.

Each account is individual. A plan applies only to the account that bought it, and the scans, findings and agent data in an account are visible only to that account.

Acceptable use

You may only scan MCP servers and repositories you own or have permission to scan. You may not use the service to attack, disrupt, or gain unauthorized access to systems you don't control. You may not attempt to circumvent usage limits through multiple accounts, automated signup, or similar abuse.

Billing

Paid plans use Razorpay Standard Checkout and are purchased one cycle at a time. Aevrin does not automatically renew or automatically charge another cycle. Paid access remains active through the displayed paid-until date. Pro is bought monthly or annually. Charges are not prorated: buying another cycle takes effect immediately and extends the paid-until date by one full cycle.

Refunds

Payments are non-refundable except where required by applicable law, including India's Consumer Protection Act, 2019 and its e-commerce rules. Duplicate charges, and payments for access that was never delivered, are always refunded. The full terms, including how to request a refund, are in the Refund and Cancellation Policy.

Scan results and limitations

Results reflect only the checks that complete. Partial and failed scans can miss vulnerabilities, and scanner findings may include false positives. You remain responsible for reviewing the evidence and deciding whether to install or operate a target.

No warranty

Aevrin is a security scanning aid, not a guarantee. A clean scan does not mean an MCP server is safe to use; it means the specific checks we ran didn't find a problem. The service is provided as-is, without warranty of any kind.

Limitation of liability

To the maximum extent permitted by law, Aevrin's total liability arising from or related to the service is limited to the amount you paid us in the 12 months before the claim arose. Aevrin is not liable for indirect, incidental, or consequential damages, including losses from vulnerabilities a scan did not catch. Nothing here limits liability that cannot be limited under applicable law.

Changes

We may update these terms as the product changes. Material changes will be reflected here with an updated date.

Credits

Icons on this site are from Solar by 480 Design, licensed under CC BY 4.0.