The agent-native cloud infrastructure platform for production-grade AWS

Ravion does everything you'd expect from a Terraform automation platform: remote plans and applies, managed state, versioned resource history, drift detection, approval gates, and policy checks. The difference is that modules are the product. Each module also carries its builds, deploys, rollbacks, logs, and metrics, so developers get a self-service interface instead of a pile of HCL and a run queue. Learn about stacks

A PaaS hosts your app on their infrastructure, and you're capped by whatever they choose to expose. Ravion provisions everything in your own AWS account, so you get dedicated infrastructure, direct cloud pricing, easier compliance, and no wall to hit as you scale. The developer experience feels like a PaaS. The ceiling is the entire cloud, and there's no painful migration waiting for you later. Read our product philosophy

No. Most teams start with our standard library of production-ready modules, including VPCs, ECS and EC2 services, RDS, Aurora, ElastiCache, Lambda, and static sites, configured through typed inputs in the dashboard or a YAML file. When you need something custom, bring your own Terraform: run it through Ravion directly, or package it into a reusable module with typed inputs your team can self-serve. Explore the standard library

Not today. Pipelines run Terraform and OpenTofu, so CloudFormation, AWS CDK, Pulumi, CDK for Terraform, SST, and the Serverless Framework aren't supported yet, though support for more IaC tools is planned. If you have infrastructure in one of those tools already, you don't have to rewrite it to get started: import the existing resources into a Ravion module's Terraform state and leave the rest where it is. Read the import guide

Our standard library modules and pipeline runners are AWS today. Runners are temporary EC2 instances in your own account, and the native modules cover the common AWS building blocks. That doesn't limit you to AWS: authenticate a runner with credentials for any other provider, passed in as environment variables from Secrets Manager or Parameter Store, and you can manage GCP, Azure, Cloudflare, Datadog, or anything else Terraform supports, with the same plans, approvals, and state history. See the Terraform stack module

Yes. Native GCP and Azure modules and pipeline runners are coming. Until then, you can provision both clouds today by authenticating an AWS pipeline runner with the appropriate provider credentials and running your Terraform or OpenTofu through Ravion.

No. Ravion runs open source Terraform and OpenTofu in your own cloud account, and our standard library modules are open source too. Fork them, lock them down, or replace them entirely with your own. Your projects, environments, modules, and pipelines are config as code you can pull down as plain YAML at any time, and Terraform state lives in your own S3 backend or in our managed backend, which you can also pull whenever you want. View the open source modules

Your infrastructure keeps running, because it was never ours. Every resource lives in your cloud account, provisioned by open source Terraform, using state you can pull with one command. Worst case, you keep the state and the modules and run plans and applies yourself, exactly as if you'd always managed them by hand. You'd lose our dashboard and pipelines, not your infrastructure. Read about state storage

Your apps stay up. No requests, traffic, or customer data pass through Ravion. It's a control plane that automates your cloud account, so an outage on our side never touches your running services, databases, or load balancers. New builds, deploys, and Terraform runs wait until we're back, and your cloud provider's console and CLI are always there as an escape hatch. See how Ravion works

You connect an account by creating an IAM role with a CloudFormation stack we walk you through, and deleting that stack permanently revokes our access. Terraform, builds, and pipelines execute on temporary EC2 instances inside your own account, so your source code never enters our control plane and runtime secrets stay in AWS Secrets Manager or Parameter Store. Read about security

Yes. Existing resources can be imported into a Ravion module's Terraform state without recreating or destroying anything, whether you're importing into a standard library module or into your own Terraform. Coming from Flightcontrol, there's a guided migration you can run service by service, since both products share the same organization and can run side by side. Read the import guide

It's a design goal. Agents work through the CLI, API, and MCP server against config as code: dry run first, review the diff, then apply. Every infrastructure change runs through a pipeline with a visible Terraform plan and an approval gate you control, and agents pick from the modules your platform team approved instead of hand-writing IaC. They get speed inside boundaries you set. Start here with an agent

Our company was founded in 2021 and brought Flightcontrol to profitability. Development on Ravion began in June 2025, and Ravion has had production users since spring 2026. Visit Flightcontrol

Ravion is free for individuals, with 1 user, up to 50 modules, and 2,000 pipeline minutes. Team is pay as you go with no platform fee and no seat pricing: modules are free, $5, or $30 per month depending on category, and 50,000 pipeline minutes are included. You always pay your cloud provider directly for the underlying infrastructure, with no markup from us. See pricing