Every Relay user is making the same decision under the same clock: free accounts close August 15, 2026, paid accounts September 14, 2026 (timeline and export checklist). There is no universal answer; the right replacement depends on which of three things your Relay workflows optimized for: reach across many apps, power in complex logic, or a human approving before anything executes.
n8n also belongs in this conversation for technical teams; it gets its own page. This one compares the three tools most Relay teams actually shortlist against flo.space, the one product on the list built around Relay's approval model.
Side by side
| Relay.app | Zapier | Make | flo.space | |
|---|---|---|---|---|
| Status | Shutting down: free ends Aug 15, paid Sep 14, 2026 | Established | Established | Early access, onboarding now |
| Human approval before actions execute The model Relay proved | Built in, first class | Optional step, off by default | Workarounds only | The default on every action |
| How automations exist | You build workflows visually | You build Zaps visually | You build scenarios visually | AI proposes prepared actions from your tools; nothing to build |
| App coverage | ~100+ apps | Thousands of apps | Thousands of apps | 14 business tools, deeply integrated |
| AI drafting with reasoning shown | AI steps | AI steps and agents | AI modules | Core of the product: draft, reasoning, source, risk level |
| Best at | Was: approachable approval-first workflows | Reach and ecosystem | Complex logic per dollar | Approval-gated business communications |
| Entry pricing | Shutting down | Free plan; from about $20/mo | Free plan; from about $9/mo | See pricing |
How to decide in one pass
Choose Zapier if reach decides it
Many apps, mature ecosystem, familiar builder, published Relay migration guide. The default choice for app-to-app plumbing, with approval as an optional add-on.
Choose Make if complexity per dollar decides it
The strongest canvas for branching, iteration, and transformation, the lowest entry price, and a white-glove Relay migration. Budget learning time, and accept that approvals are workarounds.
Choose flo.space if the approval step decides it
If your Relay workflows existed so that invoices, quotes, replies, and order updates went out only after a person said yes, that is the exact model flo.space is built on. AI prepares each action from Gmail, QuickBooks, Shopify, Stripe, and 10 other tools, shows its reasoning and risk, and waits. It is in early access, it is not a workflow builder, and outside its 14 tools you should pair it with one of the builders above. The migration guide shows what maps.
And n8n, for technical teams
Self-hosting, code steps, and full control, covered in the Relay vs n8n comparison.
The honest caveats, all in one place
Zapier and Make dwarf flo.space on app coverage, and both are proven at scale while flo.space is in early access. Make is the cheapest way to run heavy scenarios. n8n is the only one you can own outright. flo.space's advantages are narrow and real: approval is the default rather than an option, and the AI shows its reasoning and data source on every prepared action. If those two sentences describe your Relay use, it belongs on your shortlist; if not, one of the builders is the better home.
Common questions
There is no single answer. Zapier wins on app coverage, Make on complex scenarios per dollar, n8n on control and self-hosting, and flo.space on human approval as the default model. Sort your Relay workflows by which property each one needs, and it is common to end up with two tools.
No. Nothing imports Relay’s workflow JSON natively, and nothing else combines Relay’s approachability with first-class approvals across a broad app catalog. Every path is a rebuild; the export’s prompt.md files make rebuilds much faster in all four tools.
The export. Free accounts are deleted after August 15, 2026, paid after September 14, 2026, and the export takes up to 24 hours to generate with a 48-hour download window. Export this week; rebuild on your own schedule.
Yes. Early access onboarding connects your tools read only, so you can watch the approval queue propose actions alongside your live Relay workflows and compare the results before anything sends.
The approval-first option on the list
Connect your tools read only, watch the queue fill with prepared actions, and approve the first one when you trust it. That is the whole model.