Render

What Render actually is

Render is a managed application platform that sits between do-it-yourself VPS hosting and the full hyperscaler catalog. You connect a Git repository, describe the service you want — a web service, a private service, a background worker, a cron job, a static site — and Render builds, deploys, and operates it. Databases are first-class citizens too: managed PostgreSQL and a Redis-compatible Key Value store are provisioned from the same dashboard rather than bolted on through an add-on marketplace.

The company was founded in 2018 by Anurag Goel and is headquartered in San Francisco. In our view, the product is best understood as a more opinionated, more predictable successor to the old Heroku model: you give up low-level control over machines and networking, and in exchange you stop running build pipelines, TLS renewals, and patch weekends yourself.

The day-to-day workflow is the strongest part of the platform. Every push to a connected branch triggers a build and a zero-downtime deploy; every pull request can get its own preview environment with a real URL, so reviewers click through the actual change instead of reading a diff. Infrastructure as code exists in the form of render.yaml blueprints and a Terraform provider, which we consider table stakes for any team that wants to reproduce an environment rather than click it together twice.

Service types cover the usual application shapes. Web services get automatic TLS, custom domains, health checks, and HTTP/2. Background workers and cron jobs share the same instance-type table as web services, so a worker is priced identically to a web service of the same size. Private services talk to each other over Render's private network without exposing ports publicly. That coherence matters: the pricing model does not punish you for splitting one big service into four small ones.

Pricing: what you really pay

Since April 2026, Render's billing has three layers, and it is worth understanding all three before you model a monthly bill. First comes the workspace plan: Hobby at $0 per month (one member, up to 25 services, 5 GB of included bandwidth), Pro at $25 per month flat (unlimited members and services, 25 GB included, horizontal autoscaling, SOC 2 and ISO reports), Scale at $499 per month (1 TB included, SAML SSO, SCIM, HIPAA workspaces), and Enterprise at custom pricing with a dedicated account manager. Render replaced its former $19-per-member Professional plan with this flat structure and cut included bandwidth at the same time — from 100 GB to 5 GB on Hobby and from 500 GB to 25 GB on Pro.

Second is compute, billed per instance and prorated by the second. The published ladder for web services, private services, and workers runs: Starter $7 per month (512 MB RAM, 0.5 CPU), Standard $25 (2 GB, 1 CPU), Pro $85 (4 GB, 2 CPU), Pro Plus $175 (8 GB, 4 CPU), Pro Max $225 (16 GB, 4 CPU), and Pro Ultra $450 (32 GB, 8 CPU). Custom instances go up to 512 GB of RAM and 64 vCPU by sales quote. The same tier names appear across service types, which is a genuine usability win over platforms that invent a new unit for every product.

Third are the metered extras: bandwidth beyond the included allowance at $0.15 per GB, persistent SSD disks at $0.25 per GB per month, extra custom domains at $0.25 per month, extra build minutes at $5 per 1,000, and dedicated outbound IP sets at $100 per month. Managed PostgreSQL starts at $6 per month for a 256 MB instance with storage billed separately at $0.30 per GB, and scales to large dedicated instances; note that Render recently refreshed its Postgres plan lineup (the current plan names and prices differ from older reviews you will find online). Key Value instances start at $10 per month for 256 MB with a free 25 MB tier above it.

Run the arithmetic before you commit. A realistic small production stack — Pro workspace, one Standard web service, one Standard worker, one small Postgres — lands near $80 per month before bandwidth. That is not the cheapest way to run a 2 GB container, and we say so plainly. What you are buying is the removal of an ops workload, and whether that trade is favorable depends entirely on how expensive your team's attention is.

The free tier deserves a precise description because marketing copy blurs it. The Hobby plan includes free static sites and a free web service instance (512 MB, 0.1 CPU) that spins down after roughly 15 minutes of inactivity, capped at 750 instance hours per month. There is a free Postgres instance as well, but it expires after 30 days and carries no backups. Fine for prototypes and learning; we would not put anything a customer touches on it.

Performance & infrastructure

Render publishes five deployment regions: Oregon, Ohio, and Virginia in the United States, Frankfurt in Germany, and Singapore. That is a deliberately small footprint compared with the hyperscalers or edge-focused platforms, and we would like to see more European and Asian options. For most of our readers — teams serving one continent from one region — five regions is workable, but latency-sensitive global products will feel the gaps.

On the platform side, features we consider genuinely useful include horizontal autoscaling of web services (Pro plan and up), private networking between services, health-checked deploys with automatic rollback, and high-availability Postgres with read replicas on larger instance types. Persistent disks give services real block storage, and paid Postgres plans include point-in-time recovery.

Security and network controls have improved. Every hosted application gets DDoS mitigation at no charge, powered by Cloudflare's protection infrastructure and enabled by default. Dedicated outbound IPs are available on Pro workspaces and above: each set provides three static IPv4 addresses, one per availability zone in the chosen region, so allowlisting against a payment processor or a client database stops being a whack-a-mole exercise. Two caveats we keep flagging in our comparisons: those dedicated addresses are IPv4-only, and Render still does not support IPv6 for services — an open feature request that matters if your users or your upstream providers are IPv6-native. Dedicated IPs also affect outbound traffic only; inbound restrictions are handled through separate inbound IP rules.

We have not benchmarked Render ourselves — our dated sysbench/fio/iperf3 suite covers a specific set of VPS providers (see /benchmarks), and we will not manufacture numbers for platforms we have not run. What we can say from operating experience is the qualitative part: build times are predictable, deploys are zero-downtime by default, and instance performance is consistent with the published vCPU and RAM sizes rather than wildly oversubscribed.

Where it falls short

The largest gap is contractual, not technical. Render publishes no availability SLA for Hobby, Pro, or Scale customers: the standard terms contain no uptime percentage, no credit schedule, and no claim process. A contractual uptime commitment exists only in negotiated Enterprise agreements. Third-party comparisons sometimes repeat a "99.95% paid-tier SLA" figure — we could not find that commitment anywhere in Render's own terms, and we would not rely on it. If your business needs a papered availability guarantee, Render's honest position is that you must negotiate one.

Then there is the egress story. The included bandwidth allowances — 5 GB on Hobby, 25 GB on Pro — are small for anything that serves media or heavy API payloads, and the $0.15 per GB overage adds up faster than the flat workspace fee does. This is the main way a "predictable" bill becomes unpredictable, and it changed for the worse in the April 2026 pricing update.

Control-plane depth is intentionally limited. There is no GPU compute, no bare metal, no object storage of Render's own, no IPv6, and no choice of hypervisor or OS image beyond what the build system detects (or a custom Dockerfile). Networking stays simple by design: if you need fine-grained firewall rules, BGP, or multi-region active-active topologies, this is not your platform. Region count is small, the free tier sleeps on idle, and cold starts on that free tier are the usual complaint from anyone who demos a side project to a client.

Who should pick this

Our pick when the team is small, the product is a conventional web application backed by Postgres, and nobody wants to own a Kubernetes cluster. Two-to-ten person startups shipping Node, Python, Go, Ruby, or Elixir services get most of the value: Git-based deploys, preview environments per pull request, managed databases, and a pricing model you can explain to a co-founder in two minutes.

We would steer you elsewhere in three cases. If you need a contractual SLA or compliance paperwork beyond what Scale offers, the Enterprise sales process is the only route and you should compare it against AWS or Azure offers first. If your workload is bandwidth-heavy or latency-global, the egress pricing and region count will cost you. And if you want root access, custom kernels, or GPUs, a VPS or bare-metal provider — see /category/bare-metal — is simply the right tool.

For the reader comparing PaaS options directly: Render's strongest argument versus Heroku is price at any given instance size and the coherence of its product surface; versus Railway and Fly.io, it trades raw flexibility and global edge placement for a calmer, more documented operating model. That is a defensible position, and it is ours for the typical small-team web app.

Key takeaways

• Three-layer pricing since April 2026: workspace plan (Hobby $0, Pro $25, Scale $499), per-instance compute from $7 per month (Starter, 512 MB / 0.5 CPU) to $450 (Pro Ultra, 32 GB / 8 CPU), plus metered extras.
• Five regions only (Oregon, Ohio, Virginia, Frankfurt, Singapore); no IPv6, no GPUs, no bare metal.
• No public uptime SLA outside negotiated Enterprise contracts — do not believe the "99.95%" figure floating around third-party comparisons.
• Free DDoS protection on every service (Cloudflare-powered) and dedicated outbound IPv4 sets on Pro plans and above.
• Included bandwidth is thin (5 GB Hobby / 25 GB Pro) and overage is $0.15 per GB; model egress before you sign up.
• Best fit: small teams with Postgres-backed web apps who want Git deploys, preview environments, and managed operations over infrastructure control.

Sources

• https://render.com/docs/regions
• https://render.com/docs/ddos-protection
• https://render.com/docs/dedicated-ips
• https://render.com/docs
• https://render.com/pricing
• https://whichdevtool.com/tools/render
• https://tutorsbot.com/blog/render-pricing-2026-plans-cost
• https://modsignal.io/vendors/render/changes/f5d99da6-7001-4d98-a598-8a1653ebb6ee
• https://checkthat.ai/brands/render/pricing
• https://tracxn.com/d/companies/render/__areTKBDx6cbsqhJUgtno9iiOHYv6sR5gLA4saBr4k3k
• https://linkedin.com/company/renderco
• https://ontracko.com/sla/render
• https://feedback.render.com/features/p/the-ability-to-use-ipv6-on-our-web-services