PeerLM logoPeerLM

Terms of Service

Last updated: September 21, 2026

1. Agreement

These Terms of Service ("Terms") govern access to PeerLM (the "Service"), operated by PeerLM Inc. ("we," "us," or "our"). By creating an account or using the Service you agree to them. If you are using the Service for an organization, you represent that you have authority to bind it. If you do not agree, do not use the Service.

2. The Service

PeerLM monitors production LLM traffic, runs blind comparative evaluations, and produces reports and living switch / route / hold verdicts. The Service generates or replays model responses, sends them to a judge panel, and stores the resulting evidence in your workspace. A readable description of how that data is handled is at peerlm.com/data; our Privacy Policy is incorporated into these Terms.

3. Accounts

You must provide accurate information and keep credentials confidential. You are responsible for activity under your account and must notify us promptly of unauthorized use.

4. Acceptable use

You agree not to use the Service to:

  • Generate or evaluate content that is illegal
  • Attempt to bypass rate limits, plan allowances, or security measures
  • Reverse-engineer, decompile, or disassemble the Service except where the law allows
  • Resell or redistribute the Service without written permission
  • Impersonate any person or entity
  • Interfere with or disrupt the Service or its infrastructure
  • Submit secrets, credentials, or other content you do not have the right to process through us and our model providers

5. Plans, runs, and billing

Paid plans are billed monthly in advance and include a stated allowance of monitors and comparison runs, with model inference bundled. Deploy runs started through Prompt CI count against the same allowance. Run packs are an optional monthly addition that raise the monitor and run allowance for as long as they are subscribed. Runs are counted when they start. A run that fails, or that PeerLM cancels, does not count against the allowance. Unused run allowances do not roll over between billing periods. Refunds are not provided for completed runs. You may not exceed the monitors, run allowance, or other limits of your plan. Enterprise plans are billed annually under a separate order form whose terms control where they differ. Failed payments may result in suspension of paid features until the account is current.

6. Customer content

"Customer Content" means prompts, production traffic, configurations, model outputs, scores, and other materials you submit to the Service or that the Service generates for you in the course of a run.

You retain all rights in Customer Content you provide. We retain ownership of the Service, its software, and aggregated product metrics that do not identify you or include Customer Content. Model outputs are also subject to the terms of the model provider that produced them.

You grant PeerLM a worldwide, non-exclusive license to host, copy, process, transmit, and display Customer Content solely to provide and support the Service for you — including sending the relevant portions to the model providers named in the Privacy Policy so we can generate and judge. This license ends when we delete the content under the retention rules in the Privacy Policy, except for residual copies in backups until they roll off and copies you have yourself shared (for example a public report link).

No training. PeerLM will not use Customer Content to train, fine-tune, or evaluate any model that we own or operate, and will not sell Customer Content or share it with other customers. Hosted OpenRouter inference is restricted to zero-data-retention endpoints as described in section 7. Groq is a separate provider with its own data policy.

You represent that you have all rights needed to submit Customer Content, including production logs that may contain your end-users' data, and that doing so will not violate applicable law or a third-party agreement.

7. Third-party model providers

The Service calls third-party LLM providers (including OpenRouter and Groq, and through OpenRouter the model vendors behind a given endpoint). Zero data retention is enabled on the PeerLM OpenRouter account, so hosted OpenRouter calls only route to ZDR endpoints — Azure, Bedrock, and Vertex for OpenAI, Anthropic, and Google, rather than those vendors' first-party APIs. See OpenRouter's ZDR documentation. Availability, latency, and content of those providers remain outside our control. Groq is a separate provider with its own data policy. We are not responsible for provider outages.

8. Reports and sharing

You may create share links to reports. A share link is public to anyone who has it. You are responsible for what a link discloses, including prompts and responses if you enable that visibility. PeerLM may remove a share that violates these Terms.

9. Confidentiality

We will not disclose Customer Content except to the subprocessors and providers needed to run the Service, as required by law, or with your instruction (including a share link you create). We treat Customer Content as confidential information of yours.

10. Limitation of liability

THE SERVICE IS PROVIDED "AS IS" WITHOUT WARRANTIES OF ANY KIND. TO THE MAXIMUM EXTENT PERMITTED BY LAW, WE SHALL NOT BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR PUNITIVE DAMAGES, OR ANY LOSS OF PROFITS OR REVENUE, WHETHER INCURRED DIRECTLY OR INDIRECTLY.

11. Termination

We may suspend or terminate access for violation of these Terms. You may cancel from account settings. On termination your right to use the Service ends. We delete Customer Content on the schedule in the Privacy Policy, except where retention is required by law.

12. Changes

We may update these Terms. We will notify account owners of material changes by email or through the Service. Continued use after the effective date is acceptance of the updated Terms.

13. Contact

Questions: legal@peerlm.com.