Terms of Service

AgentRQ is a human-in-the-loop task management platform for AI coding agents. It allows a connected agent to create tasks, notify human operators in real-time via MCP notification channels, and receive replies — enabling human oversight of autonomous AI sessions.

Agents reach your workspace by one of two routes:

• → Directly over MCP. An agent that speaks the Model Context Protocol — Claude Code, for example — connects to the MCP server AgentRQ exposes for each workspace.
• → Through the ACP Gateway. An agent that speaks the Agent Client Protocol (ACP) — such as Gemini CLI, Cline or Cursor — connects through @agentrq/acp-gateway, open-source software you run on your own machine. Section 5 governs your use of the Gateway.

The Service is currently in public beta and is provided free of charge. Features, limits, and pricing may change as the platform matures. We will provide reasonable notice of any material changes.

You must be at least 13 years old (or the age of digital consent in your jurisdiction) to use AgentRQ. By creating an account, you represent that you meet this requirement.

You are responsible for maintaining the confidentiality of your account credentials and for all activity that occurs under your account. Notify us immediately at hi@agentrq.com if you suspect unauthorized access.

We currently authenticate users via Google OAuth2. You agree to comply with Google's terms of service in connection with that authentication flow.

You agree to use AgentRQ only for lawful purposes and in accordance with these Terms. You must not:

• ✗ Use the Service to transmit unlawful, harmful, harassing, or abusive content
• ✗ Attempt to gain unauthorized access to the Service, its servers, or connected systems
• ✗ Reverse-engineer, decompile, or disassemble any part of the Service not covered by our open-source license
• ✗ Use the Service to build a competing product without our written consent
• ✗ Abuse the free beta by artificially inflating usage to circumvent future limits

AgentRQ exposes an MCP (Model Context Protocol) server per workspace. When you connect an AI agent to that workspace — directly over MCP, or over ACP through the Gateway — that agent acts on your behalf. You are solely responsible for:

• → The actions taken by any AI agent connected to your workspace
• → The content of tasks and messages created by or on behalf of those agents
• → Ensuring connected agents comply with applicable laws and these Terms
• → Compliance with the terms and usage policies of whichever agent vendor you connect — Anthropic for Claude Code, Google for Gemini CLI, and so on for every other agent
• → The API keys, subscriptions and vendor credentials that agent uses, which remain yours to obtain and to keep secret

AgentRQ is an independent product. We are not affiliated with, endorsed by, or sponsored by any agent vendor. “Claude” and “Claude Code” are trademarks of Anthropic, PBC; “Gemini” is a trademark of Google LLC. The Agent Client Protocol is an open standard originally developed by the Zed team. All other product names are the property of their respective owners.

The ACP Gateway (@agentrq/acp-gateway) is open-source software distributed separately from the hosted Service, from its own repository at github.com/agentrq/acp-gateway and via npm. We do not operate it. You install and run it yourself, on your own machine, under the Apache License 2.0.

The Gateway launches the agent you name as a subprocess on your machine and bridges it to your AgentRQ workspace. To do that it can read and write files in the directory you start it from, carry out the tool calls your agent requests, and relay the agent's permission requests to AgentRQ for your approval. You are solely responsible for:

• → The directory you run the Gateway in, and any file the connected agent reads, writes or deletes there
• → Reviewing permission requests before you approve them — an approval you grant is an instruction to the agent to act
• → The workspace token in your .mcp.json, which grants access to your workspace and should be treated as a secret
• → Any vendor credentials the Gateway or your agent stores locally, which stay on your machine with that agent and its vendor rather than passing to the hosted Service

Because the Gateway runs outside our infrastructure and brokers software we do not publish, the disclaimers in Section 10 and the liability limits in Section 11 apply to it in full. We make no warranty that any third-party ACP agent will behave correctly, and we are not responsible for what such an agent does on your machine.

You retain ownership of all content you and your agents submit to the Service ("Your Content"), including task titles, descriptions, messages, and attachments.

By submitting Your Content, you grant AgentRQ a limited, non-exclusive, royalty-free license to store, process, and display Your Content solely to operate and improve the Service.

We do not use Your Content to train AI models. We do not sell Your Content to third parties.

Our Privacy Policy describes how we collect, use, and protect your personal data and is incorporated into these Terms by reference.

In summary: we collect only what is necessary to operate the Service (account info, workspace data, task messages). We do not sell your data. Message history is stored securely and accessible only to you and your workspace members.

The AgentRQ core platform is open source, licensed under the Apache License 2.0. You may self-host the software subject to the terms of that license.

Use of the hosted service at app.agentrq.com is governed by these Terms, regardless of whether you also use or modify the open-source code.

The ACP Gateway is licensed under the same Apache License 2.0, but is published from its own repository on its own release cadence.

The Service is provided in beta and may be subject to outages, data loss, feature changes, or discontinuation without notice. While we aim for high availability, we make no uptime guarantees during the beta period.

We reserve the right to suspend or terminate any account that violates these Terms or places unreasonable load on the Service.

You may stop using the Service at any time. You may request deletion of your account and associated data by emailing hi@agentrq.com.

We may suspend or terminate your access at any time for violation of these Terms, or if we discontinue the Service. We will make reasonable efforts to notify you in advance of any planned termination.

We may update these Terms from time to time. When we do, we will update the "Last updated" date above and, for material changes, notify registered users via email or an in-app notice.

Continued use of the Service after changes take effect constitutes your acceptance of the revised Terms.

These Terms are governed by and construed in accordance with the laws of the State of Delaware, without regard to its conflict of law provisions. Any disputes arising under these Terms shall be resolved through good-faith negotiation, or if necessary, binding arbitration or the courts of the State of Delaware.

Questions about these Terms? Contact us at: