What $300 of Vultr Credit Actually Buys You

Vultr gives new accounts $300 in free credit, valid for 30 days, via Gibil's referral. Here's how to actually use it — burn math, region picks, three workflows.

Vultr runs a referral program. New accounts signing up through this link get $300 in free credit, valid for 30 days. _{Referral link — Gibil gets a kickback that helps fund development.}

If you're already on Gibil and have only ever used Hetzner, this is the cheapest way to try the APAC regions that landed in v0.4.0. If you're new to Gibil entirely, this is a month-long evaluation budget you don't have to put on a card.

The math

$300 over 30 days is a $10/day burn budget — and the unused remainder doesn't roll over, so the goal isn't to hoard it.

The cheapest Vultr SKU Gibil ships with is vc2-2c-4gb — 2 vCPU, 4 GB RAM, 80 GB SSD, ~$0.018/hr at list price. That works out to about $0.43/day per server running 24/7. To actually use $300 in a month, you need to commit:

~23 small servers running 24/7 for the full 30 days, or
~16,000 hourly-minimum short jobs spread across the month (~550/day) — Vultr bills hourly with a one-hour minimum, so a 15-min run still costs one hour of credit, or
A single 6 vCPU / 16 GB instance (vc2-6c-16gb) running flat-out for 30 days plus a hefty test fleet alongside it.

For a solo dev evaluating Gibil casually, you'll barely scratch the credit. For a small team putting a real workload through it for a month, $300 is enough to be the whole infrastructure bill — including the workflows below.

Why you might want APAC

The five APAC regions Vultr exposes through Gibil:

Code	Region	Useful for
`nrt`	Tokyo	Japanese users, low-latency JP integrations
`icn`	Seoul	Korean users, Naver/Kakao API testing
`sgp`	Singapore	SEA region, low-latency to AU and IN
`syd`	Sydney	AU/NZ users, AWS Sydney handoff testing
`bom`	Mumbai	Indian users, regional CDN edge testing

Run gibil providers to see the full live catalog. The default region is nrt; pass --location <code> to pick another.

Three workflows that fit comfortably in the budget

1. Integration tests with a real Postgres + Redis stack

Running five containers locally — your app, Postgres, Redis, a worker, an admin UI — turns your laptop into a heater. Boot a small server, run the stack there, run the tests there.

gibil create --provider vultr --location nrt --size small --name pg-redis-test \
  --repo github.com/you/your-app --ttl 1h --json

gibil run pg-redis-test "cd /root/project && docker compose up -d && sleep 5 && pnpm test:integration"

gibil destroy pg-redis-test

A vc2-2c-4gb gives you 2 vCPUs and 4 GB of RAM for ~$0.018/hr — about a cent for the run above. Your laptop stays cool. If a test depends on shared state, you spin up a second server and run them in parallel.

2. Multi-region smoke tests for a globally deployed app

If your app is fronted by Cloudflare or an anycast LB, the only way to find out what users in Tokyo actually see is to make a request from Tokyo. Gibil's fleet flag keeps servers in one location, but you can script across regions:

for region in nrt icn sgp syd bom; do
  gibil create --provider vultr --location "$region" \
    --name "smoke-$region" --ttl 20m --json &
done
wait

for region in nrt icn sgp syd bom; do
  gibil run "smoke-$region" \
    "curl -w '%{time_total}\n' -o /dev/null -s https://your-api.example/health" --json
done

gibil destroy --all

Five servers, twenty minutes each, ~$0.09 of credit total. The output is your real per-region p50 from inside the datacenter.

3. Build pipelines that need disk and bandwidth

Vultr's small tier ships with 80 GB of SSD (Hetzner's cax11 gives you 40 GB). For build pipelines that need to check out a fat monorepo, pull large Docker images, or stage a few gigabytes of build artifacts, that headroom matters.

gibil create --provider vultr --location nrt --size small --name asset-build \
  --repo github.com/you/big-monorepo --ttl 30m --json

gibil run asset-build "cd /root/project && pnpm install && pnpm build && \
  aws s3 sync ./dist s3://your-bucket/builds/$(git rev-parse --short HEAD)/"

gibil destroy asset-build

The egress is on Vultr's network, not your home Wi-Fi. The 80 GB disk gives you room for node_modules, intermediate build output, a Docker layer cache, and the artifacts at the same time.

What you give up

Vultr's per-hour rate is higher than Hetzner's. If you're running 24/7 workloads where cost matters more than location, stick with Hetzner — that's what the cheap EU/US baseline is for. The case for Vultr is where, not how much. APAC proximity, a bigger disk on the small tier, and the 30-day head start make it worth it for the workloads above. For a bulk EU/US test runner, the math doesn't change.

The Vultr hourly minimum also means short jobs aren't free at the second mark — a 30-second job costs the same as a 50-minute one. Bundle small jobs into one server if you're doing thousands.

And the credit is use-it-or-lose-it: anything left after day 30 evaporates. If you're going to take the offer, plan on a real evaluation in that window — load tests, the integration stack from above, a multi-region smoke run — not a hoard for later.

Get started

Vultr Setup is the hand-holding walkthrough: account signup, IP whitelist (the #1 newcomer gotcha), API key in five minutes. Or non-interactively:

gibil init --provider vultr --token $VULTR_API_KEY
gibil create --provider vultr --location nrt --size small --ttl 15m

If gibil itself isn't installed yet: npm install -g gibil.

Repo · Docs · Vultr signup with $300 credit