Relay.app is no longer available. The official Relay.app shutdown notice says the service shut down in September 2026 and that all accounts, workflows, and run history were permanently deleted. Before closure, Relay stood out for putting human approval steps directly inside visual automations; former users now need to recreate those controls on another platform rather than evaluate Relay as an active product.
Relay.app review: current status
Shutdown status: Relay.app has shut down. The company’s current site states that the shutdown happened in September 2026 and that account data, workflows, and run history have been permanently deleted.
That changes the purpose of this review. Earlier 2026 comparisons can still explain what Relay did and why teams chose it, but historical feature lists and pricing are no longer purchase information. The useful questions now are what the product did well, what former users may need to reconstruct, and which live automation platform best matches the part of Relay they valued.
What happened to Relay.app?
The strongest current evidence is Relay’s own shutdown notice. It is more authoritative than older reviews that still describe the product as available because those pages were written before the closure or have not yet been updated.
The deletion notice also matters operationally. A former user should not plan a migration around retrieving a workflow definition or execution history from the old Relay account. Any reconstruction now has to start from material that still exists elsewhere, such as internal documentation, screenshots, browser history, connected-app activity logs, team knowledge, exported notes, or records created by the original automations.
That does not mean every former workflow has to be recreated exactly. Closure is also an opportunity to identify which approval gates, integrations, and AI steps were actually useful and which parts can be simplified on the replacement platform.
What Relay.app did well before shutdown
Historical 2026 reviews described Relay as a visual no-code automation platform with built-in AI steps, team collaboration, and mainstream business integrations. The product was designed for teams that wanted automation without handing every consequential action to a fully autonomous agent.
Historical workflow architecture map: Trigger → deterministic workflow steps → optional AI step → human review or approval → downstream action. Relay’s distinctive idea was not simply “use AI in a workflow.” It was making the human checkpoint a first-class part of the workflow rather than an improvised workaround.
That pattern worked well for operations where the machine could prepare work but a person still needed to confirm the final action. Historical reviews cited client communication, content review, CRM verification, and similar approval-sensitive tasks as natural fits.
Relay’s visual builder also reduced the barrier for mixed technical and non-technical teams. A business operator could follow the path of the automation without reading a codebase, while AI steps could summarize, classify, extract, or draft information inside the flow.
Human-in-the-loop was Relay’s real differentiator
Before the shutdown, the clearest reason to choose Relay was its human-in-the-loop model. Historical reviews described workflow steps that could pause execution and ask a person to approve, edit, reject, or otherwise intervene before the automation continued.
That design solved a practical problem with AI automation: some work is useful to automate up to the point where the cost of a wrong action becomes high. Drafting a reply is low risk; sending it to an important customer is higher risk. Preparing a CRM change is different from committing it. Generating a proposed social post is different from publishing it.
A replacement platform does not need to copy Relay’s interface, but it should preserve the same control principle when the workflow needs it. The critical requirement is that approval is a deliberate workflow state with a clear owner, not merely a message asking someone to check something after the action has already happened.
Historical limitations
Relay was not the broadest integration platform. Pre-shutdown reviews commonly noted that its integration library was smaller than Zapier or Make, which could matter for niche industry software and legacy systems.
The product also occupied a middle ground between conventional workflow automation and fully autonomous agents. That was useful for approval-heavy operations, but teams that wanted deep self-hosting, custom code, internal APIs, or highly technical orchestration could find n8n more appropriate.
Conversely, teams that only needed simple trigger-action connections across mainstream SaaS products could prefer Zapier’s managed connector model. Relay’s advantage was strongest when human participation was intentionally part of the process.
How to reconstruct a former Relay workflow
Migration reconstruction checklist: because Relay says workflows and run history were permanently deleted, rebuild from the business process outward rather than trying to reproduce an unavailable configuration screen.
- Inventory the trigger. Identify what started the old workflow: a form, email, CRM change, schedule, webhook, or manual action.
- List the downstream systems. Record every app that received or supplied data, including the fields that mattered.
- Recreate the decision points. Separate fixed conditions from steps where AI interpretation was used.
- Identify every approval gate. Note who reviewed the work, what they could edit, and what happened after approval or rejection.
- Recover evidence outside Relay. Look for surviving process documents, screenshots, connected-app logs, generated records, team notes, or messages that reveal how the automation behaved.
- Review old credentials and connections. As a security precaution, check connected services for old OAuth grants, API keys, webhooks, or service accounts and rotate or revoke anything no longer required.
- Build the smallest working replacement. Recreate the critical path first rather than rebuilding every historical branch immediately.
- Test failure behavior. Force missing data, timeouts, duplicate triggers, rejected approvals, and AI errors before the replacement goes live.
- Add logs and alerts. The new workflow should make it obvious which step failed and who owns recovery.
This approach preserves the business logic even when the original platform state is gone.
Best Relay.app alternatives now
Current alternatives matrix: the right replacement depends on why Relay was useful to you.
| If you valued Relay for… | Start with | Why |
|---|---|---|
| Simple managed SaaS automation | Zapier | Strong managed connector model and low setup friction for mainstream business apps |
| Visual branching and data transformation | Make | Visual scenarios expose routers, filters, transformations, and multi-path logic clearly |
| Self-hosting, APIs, and code-level control | n8n | Better fit for technical teams that need infrastructure and workflow ownership |
| Agent-first email, scheduling, and CRM work | Lindy | Managed agents target communication-heavy business tasks without requiring explicit flowcharts for every step |
| AI-native research, scraping, and enrichment pipelines | Gumloop | Visual AI workflows combine models, scraping, documents, and approval-oriented controls |
For mainstream connector-driven automation, see our Zapier review. For visual workflow logic, see the Make review. Technical teams can compare the self-hosted approach in our n8n review. Teams that want agent-first business assistants can read the Lindy AI review, while research-heavy AI workflows are covered in our Gumloop review.
If human approval was the reason Relay worked for your process, evaluate that capability explicitly in the replacement. A platform may support AI and integrations yet still make approvals awkward to model, assign, audit, or resume.
Final verdict
Relay.app should no longer be evaluated as a product to buy. It shut down in September 2026, and its official notice says accounts, workflows, and run history were permanently deleted.
Its most useful legacy is the operating principle behind its human-in-the-loop design: automate preparation and repetitive work, but pause before consequential actions when human judgment still matters. Former users should reconstruct the business process, preserve those approval boundaries, and choose the replacement platform according to the workflow’s real control requirements rather than trying to reproduce Relay feature for feature.
