all systems normal

open beta / 240 apps connected

The handoffs run themselves.

Relay runs the steps between your apps and only pings a person when it matters.

Start free

Watch a build, 2 minutes

free for 14 days / no card / no call with sales

Refunds over 5000

3 steps / 1 branch

100%

running

Amount over 5000

condition

Log to warehouse

postgres

Ask finance

waiting on a yes

26 min

true

else

last runs

09:41:02refunds over 5000129 ms
09:38:57weekly export2.1 s
09:31:14invoice reminder410 ms

Refund created

stripe

Relay

all systems normal

Start free

open beta / 240 apps connected

The handoffs run themselves.

Relay runs the steps between your apps and only pings a person when it matters.

Start free

Watch a build, 2 minutes

free for 14 days / no card / no call with sales

Refunds over 5000

3 steps / 1 branch

100%

running

Amount over 5000

condition

Log to warehouse

postgres

Ask finance

waiting on a yes

26 min

true

else

last runs

09:41:02refunds over 5000129 ms
09:38:57weekly export2.1 s
09:31:14invoice reminder410 ms

Refund created

stripe

connects to 240 apps, and to whatever you built yourself

slack
stripe
postgres
hubspot
notion
gmail
webhooks
233 more

content service

Most of a process is glue.

Copy this into that. Check it went through. Ask Dana if it is fine. Relay takes the glue and leaves you the decisions.

One canvas, every branch

Drop a trigger, add a condition, split the path. A step that fails at 3am is retried four times.

run 41982

finished in 129 ms

read refund0 ms
check amount0 ms
branch, over 50000 ms
ask financeon Dana

050100150200 ms

Approvals where people already are

Send a step to Slack, get a yes or a no, keep moving. Nobody learns another dashboard.

#finance

today

Relay

9:41

Refund of 6,400 for Northwind is over your limit. Send it?

amount: 6,400 USD
customer: Northwind
rule: over 5,000 needs a human

Approve

Hold

waiting 26 min

Every run is on the record

Inputs, outputs and timings for 90 days. Replay any run before you trust it with money.

runs per weekday62,000 average

[ what we are actually selling ]

The best workflow is the onenobody mentions in standup.

one tick, one run

now

reliability

Boring in the ways that matter.

Credentials, retries, rate limits, residency. The parts nobody demos, and the parts that decide the year.

One runtime, every app

Relay holds the credentials, retries and rate limits, so a step behaves the same on Stripe or on the API you wrote last week.

slack
stripe
postgres
hubspot

credentials, retries, rate limits, replay

one runtime, every connection

Runs where your data lives

Pick Frankfurt, Virginia or Sydney. Payloads and logs stay in the region you chose, and the audit trail says so in writing.

where runs executed, last 24 hours

eu-central-1 Frankfurt46%
us-east-1 Virginia38%
ap-southeast-2 Sydney16%

a payload never leaves the lane it started in

daily view

A Tuesday, seen from Relay.

The screen people actually keep open: what ran, what is queued, and who still owes an answer.

this week, run by run

next run in 12 minnext in 12 min

hands off

82%

of steps finish without a person in the loop

waiting on people

3 approvals

Longest wait: 26 minutes, in #finance

failures and recoveries

recovered
failed hard

where it ran

Every run stays in the region that started it

case studies

Numbers we did not have to explain away.

Counted across every account, including the ones that churned. An honest average, not a best case.

volume
62,000
runs handled on an average weekday
up 18% since March
We killed four scripts nobody wanted to own and a spreadsheet that had been lying to us.
Marta Silvanihead of operations, Fielder
time back
11
hours a week given back per person
measured across 1,400 teams
Finance stopped asking me for the weekly export. It just lands in the channel now.
Owen Brandtdata lead, Cassette
first attempt
99%
of runs finish on the first attempt
the other 1% retried and passed
What sold my team was the log. When something breaks we see the exact payload.
Tom Ekelundcto, Bright Coast

pricing

You pay for runs, not for seats.

Invite the whole company on any plan. Nobody should ask permission to automate their own job.

estimate

How many runs do you expect a month?

your plan

Team

$89a month

30,000 runs included

Start free

Starter
$29
a month
- 3,000 runs a month
- 10 workflows
- 7 days of run history
- email support
Start free
Start free
Team
$89
a month
most picked
- 30,000 runs a month
- unlimited workflows
- 90 days of run history
- Slack approvals and code steps
Start free
Start free
Scale
$290
a month
- 300,000 runs a month
- single sign on and audit log
- 1 year of run history
- named engineer, 4 hour response
Talk to us
Talk to us

The handoffs run themselves.