Case study: A logistics startup replaced 4 internal tools with one Vibely build
Routely consolidated Retool, Airtable, Zapier, and a homegrown Node service into a single Vibely-built operations console — and dropped their stack cost by 71%.
Routely is a Series A logistics startup that helps mid-market shippers route freight across rail and trucking. By late 2025, their internal operations team was running on a stack the team called "the spaghetti": Retool dashboards, an Airtable system of record, a forest of Zapier automations, and a homegrown Node microservice that nobody on the current team had written.
The pain
- Cost: $4,800/month combined Retool + Airtable + Zapier + microservice hosting.
- Reliability: One outage every two weeks, almost always a Zap that broke silently.
- Onboarding: Four tools to learn for every new operations hire.
- Customization: Any new field meant changes in three places.
The plan
Routely's head of platform, Rahul, ran a two-week spike: rebuild the operations console as a single Vibely web app. The success criterion was feature parity, not lower cost.
How it went
Day 1-2 — Schema
Rahul described the existing data model in plain English: shipments, lanes, carriers, rate quotes, dispatch events, exceptions. Vibely generated the Supabase schema, the migrations, and a seed script.
Day 3-5 — Read views
Lists, filters, search, exports. The kind of thing Retool is great at, but with full type safety and zero per-seat licensing.
Day 6-9 — Write flows
Dispatch a shipment, raise an exception, override a rate. Each flow shipped as a small, reviewable set of turns.
Day 10-12 — Replacing the Zaps
The 22 Zapier automations were rewritten as Postgres triggers and Supabase functions. Most reduced to three lines of SQL each.
Day 13-14 — Cutover
Operations team trained for a half day. Old tools turned off the following Monday.
The numbers
- Stack cost: $4,800/mo → $1,400/mo (Supabase + Vibely subscription). 71% reduction.
- Outages: Bi-weekly → zero in the eight weeks since cutover.
- New-hire onboarding: ~3 days → ~half a day.
- Time to ship a new field end-to-end: 2-4 days → under an hour.
What surprised them
"We expected feature parity. We got better than feature parity, because the AI suggested a few small UX upgrades on the way — like inline carrier search where Retool had a modal — that we never would have prioritized." — Rahul
What did not work
One Retool feature — a complex pivot table with cross-tab calculations — was easier in Retool than in any custom build. Routely kept that one Retool view alongside the Vibely app for two months until they shipped a native equivalent.
What other ops-heavy startups should take from this
- Internal tools are the highest-ROI place to use AI builders. Speed and cost both win.
- Replace the connective tissue first. Zaps and orchestration scripts are the biggest source of silent breakage.
- Keep one or two things. Best-in-class tools that genuinely outperform a custom build are worth keeping.
- Train the operators, not just the engineers. The biggest savings show up when ops staff can request and ship small changes themselves.
Build it with Vibely
Describe what you want. Watch a working preview appear in seconds.
Start a project →
