Critical
Silent failure
0 executions in 7 days on an active automation
The trigger stopped firing but the scenario still shows active. Runmend catches it the next time we sync.
Built for agencies running client automations onMaken8n
A client workflow just failed.
Runmend noticed — knows why.
Every 15 minutes Runmend re-audits every client's Make.com and n8n workspace, catches silent failures across six rules, and asks Claude to write the post-mortem before you open the tab.
Free forever on 1 profile · Claude Sonnet diagnostics included
app.runmend.com/app/acme/profiles
0health
Workspace pulse
Acme Automations
0 profiles · 0 automations monitored
0 critical0 openSynced 2m ago
Coastal Content AgencyMake.com
47 automations · last sync 2m ago
322
34
Webhook failure — client onboarding stopped 3 days ago
GreenLeaf CommerceMake.com
23 automations · last sync 2m ago
231
62
Google OAuth token expires in 6 days
Creator Stackn8n
31 automations · last sync 2m ago
23
78
InfraFlow DevOpsn8n
18 automations · last sync 2m ago
11
91
Claude wrote a new diagnostic report 2 minutes ago
Six detectors, always on
The failure modes Make and n8n don't surface
Every 15 minutes Runmend re-runs six rules across every connected workspace. Each rule has its own threshold, severity, and remediation hint.
Critical
High error rate
More than 30% failures in the last 24 hours
Downstream API flaky, auth expired, or data shape changed — we surface it with the exact failure count.
Warning
Error spike
Error rate doubled vs. 7-day average
Something changed today that wasn't broken yesterday. We flag the baseline shift before it becomes the new normal.
Critical
Consecutive failures
5+ failed runs in a row
A streak of red is louder than a noisy error rate. We flag the exact streak length and the latest error message.
Info
Zombie automation
Active but 0 executions in 30 days
0 runs · 30d
Left over from a migration or a forgotten test. Still counts against quota, still exposes credentials.
Warning
Credential expiration
Token expires within 7 days
google_oauth
expires in 6dOAuth tokens and API keys that are about to go stale — caught before the next run silently fails.
Pre-flight Check · Reliability Suite
Catch bad output before your client does
Detectors catch what already broke. Pre-flight Check replays real test inputs through a workflow and asserts the result is still right — before the change ships. Seven assertion types, a Claude judge that scores quality against your own criteria, and drift tracking against a known-good baseline.
Structural
json_schema_valid
Output validates against your JSON Schema
{ id, email, plan }valid
The shape didn't drift. Catches a renamed key or a null where an object used to be, before it reaches a client.
Structural
field_present
A required field exists in the output
customer_idamountcurrency
The field a downstream step depends on is still there — not silently dropped by an upstream change.
Structural
field_matches
A field matches a regex or exact value
/^INV-\d6$/INV-004217 ✓
An invoice ID still looks like an invoice ID. Exact-match or pattern, your choice per field.
Structural
field_in_set
A field's value is one of an allowed set
activetrialingpast_duecanceled
Status came back as one of the values your CRM accepts — not a new enum the model invented.
Operational
latency_under_ms
Execution finished under a time budget
1.9sbudget 3s
A scenario that used to run in 2s now takes 40s. The output may be fine; the SLA isn't.
Operational
cost_under_cents
Token cost stayed under a cents budget
0.4¢ / run
cap 5¢A prompt tweak that doubled spend per run gets caught here, not on the invoice.
n8n only — Make AI-node cost isn't exposed by the platform API yet.
Semantic
llm_judge
Claude scores the output against your criterion
0.92
confidence 0.88 · “tone matches the brand voice, all fields populated”
For output that's valid but not good. Claude returns a score, a confidence, and its reasoning — optionally diffed against a baseline run.
Baseline drift, not just pass/fail
Every run is diffed against a known-good baseline. Once a scenario has 20+ inputs, Runmend reports a drift % — so you catch slow regressions, not only hard failures.
Pre-flight Check is on Pro — 5 scenarios · 50 runs/mo — and Agency — 50 scenarios · 500 runs/mo.
AI diagnostic report
Runmend writes the post-mortem before you open the tab
After every sync, Runmend feeds execution stats, error patterns, and open issues into Claude and gets back a structured report you can forward to a client in one click.
Diagnostic — Acme Automations
generated 2 min ago · Claude Sonnet
Overall health
Most dangerous issue
Recommendations
- Re-test the onboarding webhook from the Make.com scenario editor and watch for the expected payload.
- Rotate the Google OAuth token on GreenLeaf Commerce before it expires in 6 days.
- Archive the 4 zombie automations on Coastal — they have not run in 30+ days and are holding stale credentials.
Generated from 247 executionsClaude Sonnet
Inside the app
Every surface built for triage, not dashboards
Worst-first rows. Real health deltas. Channel-aware alerts. No vanity metrics.
coastal/
Needs attention · 2Criticals · 5Stale sync · 1Healthy · 2ProfilePlatformHealth · 14dIssuesLast sync
Coastal Content Agency
Content agency · 12 seats
3418
3c2w2i
2 minGreenLeaf Commerce
Ecommerce · 6 seats
624
2c3w1i
12 minCreator Stack
Media · 3 seats
783
2w3i
1hInfraFlow DevOps
DevOps · 2 seats
911
1w1i
4 minStylized preview · real surfaces ship today
Make.com
us1 · eu1 · eu2
sync · every 15 min
n8n
self-hosted
One key per platform. Runmend never stores credentials in plaintext — AES-256-GCM at rest, read only when a sync fires.
Built to catch what Make and n8n don't surface
0%
of automation users have had a workflow fail silently
without any notification from the platform
0.0 days
average time to notice a broken automation
when no monitoring is in place
$0
average revenue lost per silent failure
in unbilled work and missed orders
How it works
From blind to briefed — and bad changes caught before they ship
01
Connect
under 60 seconds
Drop in a Make.com API key or an n8n instance URL + key. We store it encrypted with AES-256-GCM and never write it to logs.
02
Detect
every 15 minutes, automatically
Six rules run on every sync — silent failure, high error rate, error spike, consecutive failures, zombie automation, credential expiration.
03
Diagnose
one click from a client-ready report
Claude reads the execution stats, open issues and error patterns and writes a three-part diagnostic: overall health, the most dangerous issue, and what to do next.
04
Prevent
on demand, before you ship
Before a change goes live, Pre-flight Check replays real inputs through the workflow and asserts schema, fields, latency, token cost, and Claude-judged quality. Drift against a baseline is flagged automatically.
From our pilot agencies
Built with the people who actually live in Make.com at 2am
Three agencies have been running Runmend against real client workspaces while we built it. Here is what they told us mattered.
“We had a Make scenario quietly returning 200s with empty payloads for nine days before a client noticed. Runmend would have flagged it the same morning. That is the whole pitch.”
“Half our incidents are expired Google or Stripe creds nobody remembers connecting. We stopped finding out from clients.”
“The Claude diagnostic doesn't just say 'error rate is 12%' — it tells the junior on call which scenario to open first and why. Cut our triage time in half.”
Quotes paraphrased from pilot conversations. Names anonymized at request.
Pricing
Start free. Scale as your client book grows.
Every plan runs all six detectors, AI post-mortems, and multi-zone coverage. Pro and up add Pre-flight Check and richer alerting — higher tiers just monitor more clients, more often.
Free
Monitor one workspace.
$0/mo
- 1 profile
- 1 sync per day
- 3 AI reports per month
- All 6 detectors
Starter
For one agency, a handful of clients.
$19/mo
- 5 profiles
- 4 syncs per day
- 20 AI reports per month
- Email + in-app alerts
Most popular
Pro
Scales with your whole client book.
$49/mo
- 25 profiles
- 24 syncs per day (every hour)
- Unlimited AI reports
- Per-profile alert routing
- Pre-flight Check · 5 scenarios, 50 runs/mo
Agency
For agencies with 50–100 client automations.
$149/mo
- 100 profiles
- 96 syncs per day (every 15 min)
- Unlimited AI reports
- Slack + custom alert rules
- Pre-flight Check · 50 scenarios, 500 runs/mo
- Priority support
Lifetime
Lifetime
Pro tier, forever. Limited seats.
$99once
- Pro limits, permanently
- Pre-flight Check · Pro limits
- One-time payment
- VAT-ready invoices
- Grandfathered into future features
Limited to 20 lifetime seats
Full comparison at /pricing. Enterprise and white-label on request.
Coming soon
Benchmarks
How does your automation health compare to the median agency?
Every agency wonders if they are doing this well. Nobody can answer it because nobody can see across agencies. As Runmend audits more Make.com and n8n stacks, the aggregate becomes the answer — your numbers, in context.
Error-rate percentile
Where your average scenario sits versus the Make.com agency median.
Coverage quality
How thoroughly your client stacks have alerts and field-mapping coverage versus peers.
Time-to-detect
How fast Runmend catches your silent failures versus the cohort baseline.
Pilot agencies get benchmarks first
Benchmarks unlock once we hit pilot density across Make and n8n agencies.
Live demo
Four real audits, four different verdicts
Explore four client-shaped profiles — one critical, two warnings, one healthy. See exactly what Runmend would surface and what Claude would write.
Make.com
34
Coastal Content Agency
Marketing Agency
Full-service content agency running 47 Make.com scenarios for client onboarding, content scheduling, reporting, and billing.
47 automationsOpen audit
Make.com
62
GreenLeaf Commerce
E-commerce
Sustainable e-commerce brand running 23 Make.com scenarios for order processing, inventory sync, email marketing, and customer support.
23 automationsOpen audit
Make.com
91
Creator Stack
Digital Courses
Online course creator running 15 Make.com scenarios for student enrollment, drip content delivery, community access, and payment processing.
15 automationsOpen audit
n8n
52
InfraFlow DevOps
SaaS / DevOps
Engineering team running 31 self-hosted n8n workflows for CI/CD notifications, incident management, customer onboarding webhooks, and internal tooling.
31 automationsOpen audit
Stop firefighting. Start monitoring.
Start in under a minute. The free tier watches 1 profile and ships 3 Claude-powered diagnostics a month — no credit card, no catch.
Free plan available. No credit card required. See pricing