Acceptable Use Policy (AUP)
Effective date: 2026-08-03 Applies to: all users of dev-mux managed hubs and the dev-mux control plane.
This Acceptable Use Policy is incorporated into and forms part of the dev-mux Terms of Service (Terms of Service). Capitalized terms not defined here have the meaning given in the ToS.
1. Purpose
dev-mux provides a managed control plane for tmux + AI coding agents in an isolated, resource-capped per-tenant environment ("managed hub"). The managed hub is a general-purpose compute resource. This policy sets the boundaries of acceptable use so that one tenant's behavior cannot degrade the Service, harm third parties, or expose dev-mux to legal or infrastructure risk.
2. Prohibited uses
You may not use the Service, a managed hub, or any dev-mux-provided compute to:
- Cryptocurrency mining. Running cryptocurrency miners, proof-of-work solvers, hash-farming, or any workload whose primary purpose is to mint, mine, or transact cryptocurrency or other digital tokens is prohibited. A low-cost managed hub farming crypto is an explicit abuse case we monitor for.
- Resource abuse. Sustained resource exhaustion beyond your plan's fair-use envelope, including crypto-mining, distributed computing for hire, bulk media transcoding farms, or any workload designed to consume compute you did not pay for.
- Network abuse. Denial-of-service or amplification attacks, port scanning of third parties, spam or bulk unsolicited email, botnet command-and-control, proxy/VPN-for-hire relays, or traffic that violates a third party's rights.
- Unlawful or infringing activity. Anything illegal in your jurisdiction or ours; storing or transmitting content you have no right to; distributing malware; or facilitating harm to others.
- Circumvention. Attempting to bypass plan limits, egress caps, quotas, authentication, or the operator monitoring boundary (see the Privacy Policy); or interfering with other tenants.
3. Technical guardrails
To protect the Service and other tenants, managed hubs are subject to automated guardrails. Each of the following is in force today:
- Resource caps. Every managed hub runs under a fixed CPU and memory ceiling set by your plan and enforced by the operating system kernel. A workload that asks for more is throttled at the ceiling rather than allowed to exceed it. This is why sustained abuse cannot consume compute you did not pay for, and cannot degrade another tenant's hub — but note that it applies to your hub too: a runaway workload of yours will make your own sessions slow long before it affects anyone else.
- Tenant isolation. Managed hubs are network-isolated from one another by default. A hub cannot open a connection to another tenant's hub.
- CPU-anomaly and restart alerts. Managed hubs are monitored for sustained CPU saturation — holding the plan ceiling continuously, which is the signature of crypto-mining and similar abuse — and for repeated restarts. Anomalies raise an operator alert for review. Short bursts of heavy CPU use are normal development work and do not raise an alert.
Egress. We do not currently meter or rate-limit your outbound bandwidth, and we do not want you to plan around a cap that is not there. We reserve the right to introduce an egress cap, and to throttle or suspend a managed hub whose outbound traffic threatens the Service or third parties. We will give notice of a general egress cap before it takes effect.
These guardrails observe system-level signals only (CPU/mem/disk, restarts, uptime). Consistent with the operator monitoring boundary, they do not read your session content, terminal traffic, or the contents of your code. See the Privacy Policy.
4. Enforcement
We prefer to warn first. Where practical we will contact you before acting. For clear, ongoing, or high-risk abuse (e.g. active crypto-mining or a network attack) we may suspend or throttle a managed hub immediately to protect the Service and third parties, then notify you. Repeated or egregious violations may result in termination under the ToS.
5. Reporting
Report suspected abuse to [email protected].
Questions: [email protected]
Terms of Service · Privacy Policy · Refund Policy · Third-party notices