Relay.app Is Shutting Down. Your Business Infrastructure Shouldn't.
Relay.app deletes free accounts August 15, paid accounts September 14. If your operations run on rented automation tools, here is the risk you have been carrying.
Every business that built operations on Relay.app discovered this week that none of it belonged to them. On July 16, 2026, the company announced it is shutting down. Free accounts delete August 15 at 23:59 PT (relay.app). Paid accounts follow September 14 (Relay.app Docs). The exit package: a data export window and a deadline.
If your business runs client onboarding sequences, internal approval routing, or data handoffs through Relay.app, you have hours or weeks to rebuild something your operations depend on. That is the immediate problem. The structural one is why this situation keeps arising.
---
The case for tools like Relay.app was always the same: connect your apps, automate the handoffs, save time. It worked. Thousands of businesses built real workflows — sequences that routed client intake forms to CRMs, triggered billing actions, sent internal notifications. Work that previously required coordination now ran automatically.
Monthly subscription pricing made the model easy to justify. What it obscured: none of those businesses owned any of what they built. The workflows lived on Relay.app's servers, accessible through Relay.app's interface, dependent on Relay.app's continued operation. When Relay.app made a business decision to wind down, every workflow became a countdown timer.
---
Relay.app is not an isolated case. It follows a documented pattern in the SaaS automation market.
Zapier has raised prices and restricted free tier limits multiple times since 2022. Make.com rebranded after acquisition, requiring customers to migrate their entire workflow infrastructure. Pipedream pivoted its roadmap away from small business use cases. Flowdash was acquired and deprecated. In each instance, the businesses most exposed were those that had treated vendor interfaces as their operations infrastructure.
The businesses that absorbed these disruptions with the least damage shared one characteristic: their critical automation logic existed in systems they controlled. The vendors changed; the operations continued.
---
Here is the counterargument worth addressing: automation tools like Relay.app, Make.com, and Zapier are inexpensive relative to the alternative. Building owned infrastructure requires engineering expertise most small businesses do not have in-house, and the up-front cost is not trivial. For isolated, low-stakes tasks, renting capability from a SaaS vendor is a rational choice.
The logic breaks down when those tools become load-bearing. Client onboarding sequences, billing triggers, internal approval workflows, and data routing between core systems are not peripheral automations. They are operational infrastructure. When that infrastructure depends on a vendor's continued operation, the business has transferred control of a critical function to a counterparty with no contractual obligation to maintain it.
---
There is a structural distinction between automation tools and owned infrastructure that subscription pricing tends to make invisible.
Monthly payment buys access to execution capability on a vendor's engine. Ownership of the data schema and the execution environment is not part of the transaction. Stopping payment ends the capability. A vendor shutdown means data exits on their timeline, not yours.
When the agent logic, execution environment, and data handling all live in systems you control, the business rules are yours to read, modify, and move on infrastructure you select. No vendor decision changes the terms.
For a business running a client onboarding sequence, the operational difference is significant. An onboarding process built inside an owned AI agent can be modified in an hour, audited step by step, and continues operating regardless of what any third-party vendor decides. An onboarding process built inside Relay.app became unavailable the moment Relay.app made its announcement.
---
The immediate practical priority, if you are still on Relay.app: export every workflow definition you can access. Free accounts delete August 15; paid accounts delete September 14 (relay.app). Document the logic manually where export tools fall short. Preserve the business rules, not the platform integration — those can be rebuilt on different infrastructure.
The longer question is where your business currently sits on that spectrum. Which critical workflows — client communication, billing triggers, internal coordination, data routing — run through vendor-controlled interfaces you cannot inspect, move, or guarantee will exist in six months? Whether your operations infrastructure is rented or owned is a current operational state, not a theoretical risk management question.
---
DeployLabs builds AI agent systems that run on infrastructure clients own. The agents, the data pipelines, the execution logic — all of it is yours. You can move it, modify it, shut it down, or bring it in-house. No vendor announcement changes that.
The AI Readiness Assessment maps your current workflow infrastructure, identifies where critical operations are exposed to platform risk, and produces a prioritized implementation plan for moving those operations onto owned systems. It is a two-week engagement at $2,500.
If Relay.app's shutdown has surfaced dependencies you were not tracking, book a discovery call to walk through what owned infrastructure looks like for your business.