"Infrastructure as code" started with Terraform files describing a cloud server. In 2026 the idea has spread well past that: platform teams now use it to give developers self-service access to infrastructure without a ticket queue, and some teams are even applying the same idea to dev laptops and physical robot fleets. Here are 8 real tools doing that, pulled from our catalog with actual pricing and honest trade-offs — not the vendor's own pitch.
1. env0 — a governance layer on top of the Terraform you already have
env0 sits on top of existing infrastructure-as-code tools — Terraform, OpenTofu, Pulumi, CloudFormation — and runs them through a governed pipeline with approvals, drift detection and compliance policies, instead of engineers running terraform apply from their laptop.
Pricing: starts at $1,500/month, higher tiers on request, with unlimited users included.
Good for: teams that already have real Terraform/Pulumi code and need approvals, audit trails and drift detection around it, not a rewrite of it.
Watch out for: $1,500/month is a real floor that prices out small teams, and it's a management layer, not a replacement — you still need to know the underlying IaC tool it's governing.
2. Humanitec — lets developers request infrastructure without waiting on ops
Humanitec is a platform orchestration layer: developers request cloud resources themselves through a self-service interface, while security, cost and compliance guardrails set by the ops team are enforced automatically underneath, without a manual ticket in between.
Pricing: starts around $2,199/month for 5 users, scaling to Enterprise or self-hosted on request.
Good for: mid-size to large engineering orgs where the ops team has become a bottleneck for routine infrastructure requests.
Watch out for: the starting price puts it out of reach for small teams, and adopting it means committing to Humanitec's model of how a platform should be structured, not just bolting on a feature.
3. Kratix — open-source, self-hosted, and now built for AI agents too
Kratix lets a platform team build a self-service capability once — "spin up a database," for example — and expose it through a consistent API, CLI or portal to developers, automation scripts, and now AI agents as well.
Pricing: free and open source to self-host; Enterprise and "Agentic" tiers are quote-only.
Good for: platform teams that want to own their stack fully instead of depending on a vendor's hosted control plane, and are already comfortable self-hosting.
Watch out for: self-hosting an open-source platform tool means you own the operational burden of running it too, and the paid Enterprise/Agentic tiers aren't priced publicly, so budgeting needs a sales conversation.
4. StackGen — an AI agent that handles the 3am DevOps work
StackGen is built around an AI agent, Aiden, that takes on the parts of the SRE job that eat up a night on call — incident response, generating infrastructure-as-code, observability and CI/CD tuning — across multi-cloud setups.
Pricing: not published; demo and free trial available on request.
Good for: teams curious about handing routine incident response and IaC generation to an AI agent rather than a human on call.
Watch out for: pricing opacity makes it hard to compare against Harness or Spacelift without a sales call, and letting an AI agent generate infrastructure code unattended is still a trust decision most teams will want to phase in carefully.
5. Kedge — instantly clone your app's whole environment
Kedge runs each app in its own small virtual machine that can be copied instantly, so testing a risky change or spinning up a preview environment for a teammate takes seconds instead of a full new deployment.
Pricing: roughly $15/vCPU-month, $5/GB-month memory, $0.05/GB-month storage, $0.01/GB egress, with a free tier including $5/month in credits.
Good for: teams that want disposable, instantly-cloned preview environments without the overhead of Fly.io, Railway or Vercel's deployment model.
Watch out for: granular usage-based pricing across compute, memory, storage and egress is harder to forecast than a flat plan, and it's a newer entrant competing directly with well-established platforms.
6. Spawned — provisions real cloud infrastructure across AWS, GCP and Azure
Spawned is a cloud platform for deploying actual infrastructure — not just serverless functions — to AWS, GCP or Azure, with centralized provisioning, scaling and operations management from one place.
Pricing: usage-based, tied to actual cloud resource consumption (compute, storage, bandwidth); $20/month in complimentary credits for new users, no published tiered plans.
Good for: teams that want real infrastructure control across multiple clouds instead of the vendor lock-in of a single serverless platform like Vercel.
Watch out for: no fixed-tier pricing makes cost prediction harder than a flat-rate PaaS, and multi-cloud provisioning tools are competing against very entrenched incumbents (Terraform, Pulumi) doing the same job.
7. Codify — infrastructure-as-code for your own dev machine
Codify takes the "describe it in a config file, run one command" idea and points it at your development computer instead of the cloud: list the languages, editors and apps you need, and it builds the setup automatically on a new machine.
Pricing: free (CA$0/month, 20 AI messages/week, 3 members, 3 files); Freelancer CA$5/month; Team CA$20/month (25 members, 50 files); Enterprise on request.
Good for: teams onboarding new engineers often and tired of a day-one checklist of manual installs, as a lighter alternative to Ansible or Nix for this specific job.
Watch out for: it solves a narrower problem than Ansible, Puppet, Chef or Nix — dev machine setup, not production infrastructure — so it complements those tools rather than replacing them.
8. Miru — configuration-as-code for robots and IoT fleets, not servers
Miru applies the same discipline developers use for cloud infrastructure to physical fleets: when a company has hundreds of robots or IoT devices in the field, Miru lets them push a configuration change once and roll it out across every device, instead of touching each one by hand.
Pricing: not published; likely quote-based given the enterprise IoT/robotics target market.
Good for: hardware and robotics companies managing device fleets that currently rely on manufacturer-specific tools or custom in-house scripts.
Watch out for: it's a niche category outside the usual cloud-IaC conversation, pricing isn't public, and it's competing against deeply entrenched proprietary tooling from hardware vendors.
Quick comparison
| Tool | Best for | Starting price | Scope |
|---|---|---|---|
| env0 | Governance over existing Terraform/Pulumi | $1,500/mo | IaC governance |
| Humanitec | Self-service infra requests for developers | ~$2,199/mo | Platform orchestration |
| Kratix | Self-hosted, open-source platform toolkit | Free (self-hosted) | Platform engineering |
| StackGen | AI agent for IaC + incident response | Not public | AI-driven DevOps |
| Kedge | Instantly cloned preview environments | ~$15/vCPU-mo | Hosting / previews |
| Spawned | Multi-cloud infra provisioning | Usage-based | Cloud deployment |
| Codify | Dev machine setup as config | Free / CA$5/mo | Dev environment IaC |
| Miru | Config-as-code for robot/IoT fleets | Not public | Physical fleet management |
The honest split here is between governance and self-service. If your team already writes Terraform and just needs approvals and drift detection around it, env0 is the narrower fix; if the real bottleneck is developers waiting on ops for every request, Humanitec or the self-hosted Kratix solve a bigger structural problem. Kedge and Spawned are closer to plain deployment platforms than platform engineering proper, and Codify and Miru show how far the "describe it, don't click it" idea has spread outside the cloud. None of these are a drop-in replacement for Terraform itself — they all assume infrastructure-as-code is already the starting point, not the thing you're adopting for the first time.