HEARTH HOSTING · EST. 2017 · AMS / IAD / SIN / SYD
Hearth is hosting for teams who measure their lives in milliseconds. Deploy in thirty seconds, run on dedicated NVMe across 14 regions, and sleep like it's handled.
Everything between git push and your user's screen is our job. Here is what that actually means, in five parts.
Push to Git and Hearth handles the rest — build, image, route, cache. No YAML, no pipeline archaeology, no 40-step "it works on my machine" saga.
Every plan sits on dedicated NVMe — no shared disks, no noisy neighbors stealing your IOPS at 9 a.m. The read path stays flat when the traffic spikes.
Every push gets its own environment: isolated database, real TLS, shareable link. Review code the way users will see it — not a screenshot from 2019.
TLS that renews itself, HTTP/3, image CDN, edge caching, DDoS shielding. You configure it exactly once — never.
Support is staffed by engineers, not a ticket farm — and they have read the changelog, unlike most of us at 3 a.m.
Composite of TTFB, throughput and deploy speed across 30 public benchmarks. We lose on price. We never hide the chart.
For the weekend project that gets out of hand.
For products with real traffic and real users.
For scale that scares the finance team.
Private racks, SSO/SAML, a dedicated account engineer, and an SLA your lawyers will actually like. Built for teams running hundreds of services.
Unedited · names used with permission and mild flattery
We moved 61 services over a weekend. The only incident was a coffee spill, and honestly that's more than any of us can say for last quarter.
Deploy times went from eleven minutes to eleven seconds. The team thinks we hired a better team. We just changed a DNS record.
Support replied in four minutes at 3 a.m. It was an actual human who had read the code. I keep a screenshot in my "things that shouldn't be possible" folder.
Our TTFB dropped 40% the day we switched regions. The graphs were so clean my PM cried a little. Just a little.
Nine years of self-hosting and Hearth still feels like a small magic trick. At this point I've stopped being grateful and started being suspicious.
Everything else lives in the docs — or ask the humans. Median reply: 4 minutes, 32 seconds.
Yes. White-glove migration is free on Ember and above — an engineer joins your call, snapshots everything, and cuts over during your quiet hours. Median migration time: one afternoon. Your users notice nothing, which is the whole point.
Every month you're under SLA, we credit 10% of that month per 0.01% missed — automatically, no forms, no "open a ticket about your ticket". In four years, 61 of 400,000 accounts have needed it. We keep the list. It's short and we're proud.
On dedicated NVMe in Frankfurt and Ashburn, with edge cache in all 14 regions. You pin a region at deploy time, and if you ask nicely, your data never leaves the EU. We even have a checkbox for it, because apparently we're serious people.
Bring them. DNS, TLS, HTTP/3 and image CDN configure themselves the moment your domain points at Hearth. ACME renewal is our problem, not yours — it has quietly succeeded 1.9 million times so far.
You get a heads-up at 80% of any resource. Upgrade in one click and the change is live in under a minute — zero rebuilds, zero downtime. Downgrades take effect at the next billing cycle, and nobody sends you a "sad trombone" email.
It's a median, measured live, published monthly on the status page. If a ticket ever takes longer, the on-call engineer writes the reason in our changelog. We've done it eleven times. Eleven, in four years.
Fourteen days free. No card, no sales call, no "synergy". Your first deploy is thirty seconds away — the oven's already warm.
38,412 apps deployed this month · 6,204 of them on a Friday