Relay.app Shutdown: Migration Guide 2026

Affiliate disclosure: This post contains affiliate links. If you buy through them, we may earn a commission at no extra cost to you. We only recommend tools we’d run our own business on.

What actually happened to Relay.app

Relay.app announced it is shutting down. Free accounts went dark on August 15, 2026; paying customers kept access for free through September 14, 2026, after which paid accounts and all associated data were permanently deleted (checked October 2026). If you are reading this after those dates, your Relay data is already gone — this guide is about rebuilding, not exporting.

Contents

What actually happened to Relay.app
Path 1: Make — the closest 1:1 for approvals
Path 2: n8n — maximum control, you build the gate yourself
Path 3: Lindy — the AI-native option
The migration paths, side by side
Rebuilding approval logic: the checklist nobody gives you
FAQ
Methodology
Sources

Relay.app’s real differentiator was its human-first design — approval steps, human-in-the-loop checkpoints, and collaborative runs were first-class citizens, not bolt-ons. No tool below reproduces that exact experience. What follows is an honest map of where that specific kind of workflow lives now, and what each move costs you in money, rebuild effort, and features.

Path 1: Make — the closest 1:1 for approvals

Rubber stamp and pen — approval workflow decision

Make has a native Approval app: create an approval request, send it to a person, and pause the scenario until they approve, decline, or it times out. For the most common Relay pattern — “draft the thing, a human reviews it, then it goes out” — this is the closest drop-in replacement.

Why it fits Relay refugees: the approval logic is visual, auditable, and the approver interacts by email or link — no Make account needed for the person clicking approve. Make’s free plan includes 1,000 credits/month, and Core starts at $9/month annual billing with 10,000 credits (checked October 2026). See our full cost breakdown in Make credits explained and the three-way comparison in n8n vs Make vs Zapier cost math.

What you lose: Relay’s approvals felt like a collaboration surface — context, comments, and the work itself in one place. Make’s approvals are transactional: a request goes out, a decision comes back. If your approval flow had rich discussion, budget for a Slack or email thread alongside.

Effort: Low to medium. Simple approve/decline gates rebuild in an afternoon. Multi-stage chains (approver A, then approver B, with conditional rework loops) take a day or two to get right.

Best for: Teams that want the fastest route back to working approval workflows without touching code. Make

Path 2: n8n — maximum control, you build the gate yourself

n8n has no built-in “approval step.” Instead, you build one from three well-worn patterns:

Wait node with webhook resume — the workflow pauses mid-run and only continues when an external callback hits a unique resume URL. Post that URL into Slack or an email with approve/reject buttons, and you have a Relay-style gate.
Wait node with form resume — n8n generates a form URL; the reviewer fills in their decision and the workflow resumes.
Human-in-the-loop via messaging — the established community pattern is a sub-workflow that posts to Slack with approve/reject buttons; the button click calls the resume webhook.

Why it fits Relay refugees: unlimited executions on self-hosting (roughly $4–6/month on a small VPS or ~$3.70/month on managed one-click hosts, checked October 2026), your data stays on your infrastructure, and complex multi-approver logic is genuinely easier to express in n8n than in any GUI builder. If your Relay workflows touched sensitive data, this is the only path that keeps that data under your roof. Cost math here: n8n vs Make vs Zapier.

What you lose: Time. Every approval surface is something you now design, build, and maintain. Relay’s polished approver experience becomes your maintenance ticket. The payoff is data ownership and logic no GUI builder can express.

Effort: Medium to high. A basic webhook-resume gate is a half-day job if you know n8n. A polished, logged, timeout-safe approval chain with audit history is a week of real work. Our n8n workflow blueprints cover the node patterns you’ll need.

Best for: Technical teams, or anyone whose approval workflows involve sensitive data or complex conditional logic. n8n

Path 3: Lindy — the AI-native option

Lindy replaces rigid approve/decline gates with AI agents that do the ambiguous work and pause for human sign-off inside Slack threads or the web app — approvals are a platform feature, not something you build.

Pricing (checked October 2026): Plus $29.99/user/month (3,000 credits), Pro $99.99/user/month (15,000 credits), Max $199.99/user/month (35,000 credits). Billing is per-user, month-to-month only — no annual discount. Credit consumption is the real variable: everyday asks cost 2–250 credits, deep work 250–1,000, big builds 1,000–2,500. When credits run out, agents pause until you top up ($10 per 1,000 credits).

Why it fits Relay refugees: If your “approval” was really “a smart human triages ambiguous stuff,” Lindy is the only option where the agent itself does the triage and only escalates decisions to you. It’s also genuinely Slack-native, so adoption doesn’t require training approvers on a new dashboard.

What you lose: Predictability. Without a published per-action credit table, you learn what a month costs by living through it — and per-seat pricing punishes teams with many occasional approvers.

Effort: Low for standard use cases (inbox triage, scheduling, drafting), high if you need deterministic, auditable approval chains — that’s fighting the tool’s design.

Best for: Founders who want an AI teammate that asks permission, not a deterministic approval pipeline.

The migration paths, side by side

	Make	n8n	Lindy
Rebuild effort	Low–medium (afternoon to 2 days)	Medium–high (half-day to a week)	Low (if you accept the agent model)
Starting cost	$0–9/mo (Core $9/mo annual, checked October 2026)	~$4–6/mo self-hosted; Cloud Starter €24/mo	$29.99/user/mo (Plus)
Approval mechanism	Native Approval app (create/approve/decline)	DIY: Wait node + webhook/form resume, or Slack-button sub-workflow	Built-in: agents pause for human sign-off in Slack/web
Execution model	Credits per module run	Unlimited on self-host; execution meter on Cloud	Credits per AI action (unpublished per-action table)
What you gain	Biggest app library, fastest rebuild, no code	Unlimited runs, data ownership, complex logic	AI does the triage; Slack-native; zero building
What you lose	Relay’s collaborative feel; per-run metering punishes complex scenarios	You maintain everything; no native approval UI	Cost predictability; per-seat pricing; deterministic audit trails
Honest fit	Best 1:1 for standard approval gates	Best for sensitive data + complex logic	Best for “AI asks me first” delegation

Rebuilding approval logic: the checklist nobody gives you

Whether you pick Make, n8n, or Lindy, the rebuild succeeds or fails on the same five questions. Answer them before you build:

Inventory your gates. Count every approval step in your Relay workflows. Most teams find 3–5 approval patterns repeated everywhere — rebuild the patterns, not the workflows.
Name the approvers and their fallback. Set a timeout on every gate, and define what happens when nobody answers. Never auto-approve on timeout — a timed-out approval should fail closed (stay pending or reject), not silently pass.
Carry the decision payload. The reviewer must see the exact item under review — the final copy, the exact amount, the exact file — not a summary.
Log every decision. Approver, timestamp, decision, what they saw. When someone asks “who approved this?” three months from now, you need the answer in ten seconds.
Test the reject path. Everyone tests approve. Almost nobody tests reject, timeout, and “approver is on vacation.” Test all three before you call the migration done.

FAQ

Is there a tool that replicates Relay.app exactly?

No. Relay’s human-first collaboration surface was its differentiator, and no migration target reproduces it 1:1. Make gets closest for standard approval gates; n8n gives you the parts to build something better for your specific case; Lindy replaces the model with AI agents that ask permission.

Can I still export my Relay.app data?

The shutdown dates were August 15, 2026 (free) and September 14, 2026 (paid), after which accounts and data were permanently deleted. If you missed the window, the data is gone.

Which option is cheapest for a solo founder with 2–3 approval workflows?

n8n self-hosted on a ~$4–6/month VPS is cheapest in raw dollars, but only if your time is free. Make’s free plan (1,000 credits/month) covers a handful of low-volume approval flows for zero dollars and zero maintenance. Lindy starts at $29.99/month per user — the most expensive of the three for simple gates.

Do I need to know how to code for the n8n path?

Not for the basic Wait-node approval gate — it’s nodes and expressions. But you need comfort with webhooks, Docker basics for self-hosting, and debugging when the resume callback doesn’t arrive. If that sentence made you tired, pick Make.

What happens to my annual Relay.app subscription payment?

Per the shutdown announcement, paid subscriptions billed through Paddle were cancelled automatically and annual customers received prorated refunds. If you never saw a refund, check your email from [email protected] before Relay’s support channels go dark.

AWL verdict: For most Relay refugees, Make is the right first move — the native Approval app rebuilds standard gates fastest, and the free tier buys you time to decide. Choose n8n if your approvals touch sensitive data or need logic no GUI can express, and accept that you’re trading polish for control. Choose Lindy only if what you actually want is an AI teammate that asks permission, not a deterministic approval pipeline.

Methodology

Relay.app’s shutdown dates come from the company’s own shutdown announcement page, read in October 2026. Lindy pricing comes from multiple September–October 2026 pricing reviews that cross-check each other (Plus $29.99 / Pro $99.99 / Max $199.99 per user per month, 3,000 / 15,000 / 35,000 credits, month-to-month billing only, $10 per 1,000 top-up credits). Make pricing (Core from $9/month annual, 10,000 credits) and n8n Cloud pricing (Starter €24/month, 2,500 executions) are from pricing comparisons verified September–October 2026. Effort estimates assume a solo operator rebuilding Relay’s approval and human-in-the-loop workflows specifically, using the documented patterns for each platform (Make’s native Approval app, n8n’s Wait-node webhook/form resume and Slack-button sub-workflow patterns).