With Relay closing (free accounts August 15, 2026; paid September 14, 2026; details here), Make is a serious candidate: a visual scenario builder with deep logic, thousands of app connections, and the lowest entry price of the major builders. Make has even published a dedicated Relay.app migration offer with hands-on help.
The honest difference is philosophy. Relay was built so a person could sit inside the workflow and approve steps. Make is built for powerful unattended scenarios; human approval is possible with workarounds but it is not a native concept, and Make's learning curve is the steepest of the three recommended tools.
Side by side
| Relay.app | Make | |
|---|---|---|
| Status | Shutting down: free ends Aug 15, paid Sep 14, 2026 | Established, actively developed |
| Human approval before actions execute | Built in, first class | No native approval step; workarounds only |
| How you build | Visual workflow builder, approachable | Visual scenario canvas; very powerful, steeper learning curve |
| App coverage | ~100+ apps | Thousands of apps and modules |
| AI steps | AI drafting and reasoning steps | AI modules and agent features |
| Pricing model | Was per-workspace with credits | Free plan; paid from about $9/mo, billed by operations |
| Migration path | Export: prompt.md + workflow.json per workflow | White-glove migration offer for Relay customers |
Who Make is for
Pick Make if your Relay workflows were genuinely complex: branching, iteration, data transformation across many apps. Operations-based pricing tends to be the cheapest way to run heavy multi-step scenarios, and the white-glove Relay migration removes most of the rebuilding burden. Technical comfort helps; Make rewards people willing to learn its canvas.
What you give up from Relay
Two things. First, approachability: Relay was designed to be picked up in an afternoon, and Make is not. Second, and more fundamentally, the human-in-the-loop model. Make scenarios run unattended; approvals mean bolting together pauses, webhooks, or email hacks that you then own forever. If the approval step was incidental to your Relay use, that cost is small. If it was the point, it is the whole decision.
Where flo.space fits in this decision
flo.space is the narrow, approval-first option: AI prepares invoices, replies, quotes, and order updates from 14 business tools, Gmail, QuickBooks, Shopify, and Stripe among them, and every action waits in a queue for a person to approve. It deliberately is not a scenario builder. Relay teams often split: Make for the complex unattended pipelines, an approval-first tool for anything customer-facing. See the full comparison for where each fits.
Common questions
Not natively. You can approximate one with pauses, webhooks, or email confirmations, but approval is not a first-class concept in Make the way it was in Relay. Evaluate that workaround for your specific workflow before committing anything sensitive to it.
Make has published a Relay.app migration page offering hands-on (white-glove) help moving workflows over. Combined with Relay’s exported prompt.md files, most straightforward workflows can be rebuilt with assistance.
Often, yes. Make’s entry plans start around $9 per month billed annually, metered by operations. Complex scenarios consume operations quickly, so compare against your actual Relay run volume rather than the sticker price.
Not necessarily. A common outcome is splitting: unattended data plumbing in Make or Zapier, and approval-gated customer-facing actions in a human-in-the-loop tool. Migrating workflow by workflow is normal.
For the workflows that need a human yes
flo.space drafts the customer-facing work from your tools and holds it for approval. Join early access and keep the model Relay proved.