Terms of Service

1. Acceptance of Terms

These Terms of Service (“Terms”) govern your access to and use of Mrrlin, including our web application, local Model Context Protocol (MCP) server daemon, client libraries, and hosted coordination services (collectively, the “Service”), operated by Mrrlin (“Mrrlin”, “we”, “us”, or “our”).

By creating an account, pairing a local bridge, downloading software, or accessing the Service, you agree to be bound by these Terms. If you are entering into these Terms on behalf of an organization, you represent that you have the legal authority to bind that entity.

2. The Local-First Architecture: Local Compute and Local Tools

Mrrlin is architected as a local-first engineering platform. You explicitly acknowledge and agree to the following operational boundaries:

• Local Agent Process Execution: Autonomous agents (including Anthropic Claude Code, OpenAI Codex, and Google Antigravity) are spawned as local child processes directly on your workstation or private server using your locally installed command-line utilities.
• Filesystem Access & Operating System Permissions: Agent commands, file modifications, test runs, and git operations are rooted in your configured working directory (cwd) and execute under your local operating system user permissions. While operations are initialized in your workspace, autonomous shell commands execute with the privileges of your local user account; the working directory is not a security sandbox or container isolation boundary.
• No Cloud Code Mirroring: Mrrlin cloud servers do not mirror your local disk, do not clone your private repositories, and do not execute shell commands or builds on your behalf.

3. Bring Your Own Keys (BYOK) and External Model Provider Processing

Mrrlin operates on a Bring-Your-Own-Key model for AI foundation model execution:

• Local Storage of Credentials: Your API keys (such as ANTHROPIC_API_KEY, OPENAI_API_KEY, or provider tokens) are maintained solely in your local environment, local configuration files (e.g., ~/.claude.json or ~/.codex/config.json), or local state stores. Mrrlin servers never receive, store, proxy, or log your AI provider credentials.
• Direct Provider Model Inference: When a local agent CLI runs, it sends prompts, relevant source excerpts, and conversational context directly from your machine to your chosen AI model provider’s cloud API (e.g. Anthropic, OpenAI, or Google) under your personal or organization account. Mrrlin does not intermediate this model traffic; data processed during inference is governed by your independent terms and privacy agreements with those AI providers.
• Direct Provider Billing: Model token costs incurred through your local agent CLI runs are billed directly to you by your respective model providers according to your agreements with them.
• Third-Party Trademarks & Platform Independence: Claude is a trademark of Anthropic PBC. OpenAI and Codex are trademarks of OpenAI, LLC. Antigravity is a trademark of Google LLC. Mrrlin is an independent platform and is not affiliated with, endorsed by, sponsored by, or partnered with Anthropic, OpenAI, or Google. Your use of external AI models remains governed solely by your separate agreements with those respective providers.

4. Role of the Hosted Cloud Service

The hosted Mrrlin web application and cloud APIs provide coordination, state synchronization, and team collaboration services:

• Managing Director turn state, task queues, Kanban columns, and cycle planning metadata.
• Facilitating real-time communication between your local MCP bridge and your web browser session via encrypted relays.
• Managing account authentication, project workspace memberships, subscription tiers, and billing entitlements.

5. Intellectual Property and Code Ownership

You Own Your Code: We claim zero ownership, copyright, or intellectual property rights over the source code, prompts, deliverables, specifications, documentation, and diffs produced by you or your agents while using Mrrlin.

As between you and Mrrlin, all outputs, modifications, and software created through local agent executions are your exclusive property. Mrrlin does not use your private code or prompts to train general machine learning models.

6. Operator Autonomy, Safety Controls, and Human Oversight

Autonomous coding agents perform real-world file edits, command execution, and code commits:

• Permission Modes: You select the permission profile under which local agents operate (including read-only planning, edit acceptance, or autonomous permission bypass). You are solely responsible for the permission modes you authorize.
• Review Responsibility: Because AI models can generate syntactically plausible yet incorrect, vulnerable, or unexpected code, you are responsible for testing, verifying, and reviewing all code changes, diffs, and commits before merging or deploying them into production.
• High-Risk Operations & Command Execution Disclaimer: You explicitly acknowledge that autonomous coding agents execute real operating system commands, package installations, and file mutations directly on your workstation or private server under your user account permissions. Mrrlin does not warrant the safety, harmlessness, or correctness of generated commands, scripts, or patches. You assume all risk of data loss, repository corruption, dependency conflicts, infrastructure disruption, or system downtime resulting from agent execution.

7. Telemetry and Data Protection

Our telemetry practices are strictly limited to aggregated operational metrics and never collect source code, diffs, or conversational context. Please review our detailed Privacy Policy & Telemetry Notice for our technical guarantees regarding secret scrubbing (redact.ts) and sandboxed local PII gating (pii-gate.ts).

8. Subscription, Billing, and Cancellation

Certain features of the Service require a paid subscription. Billing is handled securely through our payment processor (Stripe). Subscriptions automatically renew unless canceled prior to the end of the billing cycle. You may cancel your subscription or manage payment methods at any time through the in-app billing portal.

9. Disclaimer of Warranties

Please read this section carefully.

THE SERVICE IS PROVIDED ON AN “AS IS” AND “AS AVAILABLE” BASIS, WITHOUT WARRANTIES OF ANY KIND, WHETHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NON-INFRINGEMENT. MRRLIN DOES NOT WARRANT THAT AGENT OUTPUTS WILL BE ERROR-FREE, SECURE, OR ACCURATE, OR THAT LOCAL EXECUTION WILL BE UNINTERRUPTED.

10. Limitation of Liability

TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, IN NO EVENT SHALL MRRLIN, ITS DIRECTORS, EMPLOYEES, OR PARTNERS BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR PUNITIVE DAMAGES, INCLUDING LOSS OF PROFITS, DATA, USE, GOODWILL, REPOSITORY CORRUPTION, OR OTHER INTANGIBLE LOSSES RESULTING FROM YOUR USE OF OR INABILITY TO USE THE SERVICE.

IN NO EVENT SHALL MRRLIN’S TOTAL AGGREGATE LIABILITY ARISING OUT OF OR RELATING TO THESE TERMS OR THE SERVICE, UNDER ANY LEGAL THEORY (WHETHER IN CONTRACT, TORT, STRICT LIABILITY, OR OTHERWISE), EXCEED THE GREATER OF ONE HUNDRED US DOLLARS ($100 USD) OR THE TOTAL AMOUNTS ACTUALLY PAID BY YOU TO MRRLIN FOR ACCESS TO THE SERVICE IN THE TWELVE (12) MONTHS PRECEDING THE CLAIM.

11. Indemnification

You agree to defend, indemnify, and hold harmless Mrrlin, its affiliates, directors, officers, employees, and agents from and against any third-party claims, liabilities, damages, losses, and reasonable legal fees arising out of or in any way connected with: (a) your access to or use of the Service; (b) local agent command execution, code modifications, or deliverables generated by you or your agents; (c) your violation of these Terms; or (d) your violation of any third-party rights, including intellectual property rights, data protection regulations, or third-party AI model terms.

12. Governing Law, Mandatory Arbitration, and Class Action Waiver

These Terms and any dispute or controversy arising out of or related to the Service shall be governed by and construed in accordance with the laws of the State of Delaware, United States, without regard to its conflict of law principles.

Any dispute, controversy, or claim arising out of or relating to these Terms shall be resolved exclusively through confidential, binding individual arbitration administered in accordance with standard commercial arbitration rules. You and Mrrlin explicitly agree that each party may bring claims against the other only on an individual basis and not as a plaintiff or class member in any purported class, consolidated, or representative proceeding.

13. Contact Information

For legal inquiries, terms interpretation, or compliance questions, please contact our team at: