Reproducible, hyperscalable dev environments

Full virtual machines that dynamically scale from 1 to 64 GB of memory as your workload demands it. No restarts, no reprovisioning. You just code. We handle the rest.

gradient-vm-01
live
1.2GB used
2GB allocated· 64 GB max
0 GB64 GB
VM ceiling
Usage
IDE starting
4 vCPU·Billed for 2 GB
$0.081/hr

The problem

Developer environments are broken

Despite advancements in containerization, the “it works on my machine” problem has moved from production to developer environments. Modern applications mock dozens of services locally just to start developing.

Containers fall short

Services relying on IPC, shared memory, Unix sockets, or systemd fail unpredictably in devcontainers. They break inconsistently across identical machines, making onboarding a multi-day ordeal.

Rigid machine sizes

Developer environments are memory-constrained, not CPU-constrained. Existing providers bundle CPU and memory together, forcing you to pay for 32 cores you don’t need just to get 64GB of RAM.

Expensive managed solutions

Managed cloud IDEs charge upwards of $2,000/mo per developer for a 64GB machine. The cost makes it impractical for most engineering teams to adopt standardized environments.

Painful self-hosting

Self-hosted platforms require complex Terraform modules, have sparse documentation, and take 5+ minutes to provision a single machine. Testing environments takes days of iteration.

Features

Built for modern development

Everything you need to create consistent, scalable development environments for your team.

01

Dynamic Memory Scaling

Your VM scales from 1GB to 64GB of memory in real time. No restarts, no reprovisioning. Memory grows with your workload and shrinks when you're done.

02

Decoupled CPU and RAM

Dev environments are memory-constrained, not CPU-constrained. Choose 4 vCPUs with 64GB of RAM instead of paying for 32 cores you don't need.

03

Sub-Second Provisioning

Machines provision in under a second. No waiting 5 minutes for Terraform. Iterate on your Devfile as fast as you iterate on your code.

Reproducible

Devfiles

Define your entire development environment in two files, tracked directly in Git. The gradient.yaml declares your machine spec and services. The Devfile runs setup commands like a Dockerfile, but for your full VM.

  • Install system packages, language runtimes, and tools declaratively
  • Configure databases, message queues, and background services
  • Version-controlled alongside your application code
  • Onboard new developers in minutes, not days
# Gradient Compute config
base: ubuntu-22.04

vcpu: 4
memory: 1GB..64GB

packages:
  - nodejs@20
  - python@3.12
  - postgresql@16
  - redis@7

services:
  postgres:
    port: 5432
    start: true
  redis:
    port: 6379
    start: true

setup:
  - npm install
  - npm run db:migrate

Comparison

How we stack up

We decouple CPU and memory, pass infrastructure savings to you, and provision machines in under a second.

Feature
Gradient
Codespaces
Coder
64GB Machine
$46/mo
$2,102/mo
Self-hosted
Provisioning Time
< 1 second
~30 seconds
~5 minutes
Memory Scaling
Dynamic (1-64GB)
Fixed tiers
Fixed tiers
CPU / RAM
Decoupled
Bundled
Bundled
Environment Definition
Devfiles (Git)
Devcontainers
Terraform
Full VM (systemd, IPC)
Yes
No
Yes
Coming soon

Gradient Agents

Because your environments are reproducible, we can give AI agents the same fully-initialized development environment your engineers use. Same OS, same services, same toolchain.

When issues are opened, Gradient Agents spin up your full stack, introspect databases, run test suites, and publish probable root causes to the pull request. They can also handle small feature requests and open PRs autonomously, running on our infrastructure in isolated virtual machines.

Automated triage

Agents reproduce bugs in real environments and surface root causes on your PRs.

Feature PRs

Delegate small tasks. Agents implement, test, and open pull requests from identical VMs.

Isolated VMs

Every agent runs in its own virtual machine, identical to your team’s dev environments.

Pricing

Pay only for what you use

No more paying $2,000/mo for a 64GB machine. We decouple CPU and memory so you only pay for the resources you actually consume. A comparable machine on Gradient starts at $46/mo.

Memory
$0.001/GB/hr
Compute
$0.02/vCPU/hr
64GB machine
~$46/mo
Create Free Account