The agent-native cloud infrastructure platform for production-grade AWS
Address
Change type
ID
app/alb-webto destroyLoadBalancersg-0abc12345 (alb-web)to changeSecurityGroup:80to createListenerhost=api.example.comto changeListenerRuleapp/tg-webto destroyTargetGroup
CopyDownload
Wrap
| 02:06:40.000 | info | Running build in Washington, D.C., USA (East) - iad1 | | ------------ | ---- | ----------------------------------------------------------------------------- | | 02:06:40.800 | info | Build machine configuration: 2 cores, 8 GB | | 02:06:41.600 | info | Cloning github.com/ravionhq/marketing-website (Branch: main, Commit: a1b2c3d) | | 02:06:42.400 | info | Cloning completed: 412.882ms | | 02:06:43.200 | info | Running build command: pnpm run build | | 02:06:44.000 | info | Installing dependencies... | | 02:06:44.800 | info | Resolving workspace packages... | | 02:06:45.600 | info | Compiling TypeScript sources | | 02:06:46.400 | info | Type checking with tsc... | | 02:06:47.200 | info | Bundling client with Vite | | 02:06:48.000 | info | Optimizing dependencies | | 02:06:48.800 | info | Generating static pages (12/12) | | 02:06:49.600 | info | Collecting build traces | | 02:06:50.400 | info | Build completed in 18.4s | | 02:06:51.200 | info | Uploading build outputs... | | 02:06:52.000 | info | Build outputs uploaded (24.6 MB) | | 02:06:52.800 | info | Assigning custom domains | | 02:06:53.600 | info | Provisioning TLS certificate | | 02:06:54.400 | info | Warming edge cache | | 02:06:55.200 | info | Running post-deploy health checks | | 02:06:56.000 | info | Health check passed: GET / 200 | | 02:06:56.800 | info | Health check passed: GET /api/status 200 | | 02:06:57.600 | info | Promoting deployment to production | | 02:06:58.400 | info | Invalidating CDN cache | | 02:06:59.200 | info | Notifying webhook subscribers | | 02:07:00.000 | info | Deployment ready | | 02:07:00.800 | info | Build pipeline finished in 1m 13s |
ravionhq/modules
prod.tfvars
engine
engine\_major\_version
instance\_class
staging.tfvars
engine
engine\_major\_version
instance\_class
allocated\_storage
availability.tfvars
multi\_az\_enabled
read\_replica\_count
backup\_retention\_period
deletion\_protection\_enabled
database/rds.tfvars
storage\_type
performance\_insights\_enabled
monitoring\_interval
99
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
variants:
\- id: default
name: Default
steps:
\- type: build:image
id: build-image
name: Build Image
source:
type: git
repo: https://github.com/ravionhq/app
branch: main
ref: << pipeline.input.commit >>
builder:
type: dockerfile
dockerfile: Dockerfile
context: .
\- type: build:image
id: build-web
name: Build Web
source:
type: git
repo: https://github.com/ravionhq/web
branch: main
builder:
type: nixpacks
nixpacks\_version: latest
build\_cmd: npm run build
RDS Database
Creates and manages an Amazon RDS database. Supports PostgreSQL, MySQL, MariaDB, Oracle, and SQL Server engines.
Overview
Use this module when an application needs a managed relational database in AWS. It provisions an RDS instance with subnet placement, security group access, backups, monitoring, optional read replicas, and CloudWatch alarms.
Networking
Select a VPC network first. Ravion maps the network's AWS account, region, VPC ID, and private subnet IDs into this module. Public access is disabled by default.
Credentials and security
- Master password created and stored in AWS Secrets Manager
- Storage encryption on by default, with optional customer-managed KMS key
- IAM database authentication for MySQL and PostgreSQL
| Type | rvn-rds | | ------------ | ------------------------------- | | Version | 0.4.2 | | Last updated | yesterday | | Source | ravionhq/modules / database/rds | | Monthly cost | $16.34 |
./module.yaml
Configuration is valid
99
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
inputs:
\- type: $ref:rvn-aws-network
id: network
label: VPC network
immutable: true
required: true
\- type: string
id: engine
label: Engine
default: postgres
immutable: true
required: true
\- type: string
id: instance\_class
label: Instance class
default: db.t4g.micro
required: true
\- type: number
id: allocated\_storage
label: Allocated storage (GiB)
default: 20
min: 20
max: 65536
required: true
The agent-native cloud infrastructure platform for production-grade AWS
Give developers and AI agents one governed way to provision, deploy, and operate applications in your AWS account. Move fast without building every DevOps capability yourself.
Deploy this project with ravion.com/SKILL.md
Trusted by engineering & platform teams at scale
Your product needs AWS. Your team doesn’t have time to build the platform around it.
You need more than your platform can give you
Your deployment and reliability requirements no longer fit the platform’s fixed workflows.
You don’t want to operate raw AWS yourself
Getting infrastructure running is only the start. The team still has to assemble networking, permissions, pipelines, monitoring, and day-to-day operations.
You can’t wait to build a DevOps team
The engineers who could build that system are the same engineers needed on the product. Hiring takes time, and the roadmap cannot wait.
With Ravion
Start with the AWS system
you would otherwise have to build.Provision, deploy, and operate through one connected system.
AI Agent
Step 1
Define what your application needs
Start with Ravion’s production-ready modules, add your own modules, or connect existing Terraform. Developers and agents work through structured inputs instead of assembling the common path from scratch.
3 modules
1 module
2 modules
1 module
1 module
3 modules
1 module
2 modules
1 module
2 modules
2 modules
1 module
2 modules
1 module
Step 2
Review the plan before anything changes
See exactly what Ravion will change, apply permissions and approval gates, and preserve an audit history of every action.
Add checkout v2 — tags and versioning41m ago/Git Push
Scale ECS desired count to 4LiveToday 17:32:49/Manual
Rotate RDS credentialsYest 18:32:49/Schedule
Initial infrastructure applySat 20:32:49/Manual
Address
Change type
ID
aws\_ecs\_service.apito changepayments-prod-apiaws\_s3\_bucket.app\_datato changepayments-prod-dataaws\_lb\_target\_group.apito createpayments-prod-api-v2aws\_lb\_listener\_rule.apito replacearn:aws:elasticloadbalancing:us-east-1:123456789012:listener-rule/app/payments-prod/50dc6c495c0c9188aws\_security\_group.ecsto changesg-0a1b2c3d4e5f6a7b8aws\_cloudwatch\_log\_group.api/ecs/payments-prod/apiaws\_ecs\_cluster.mainpayments-prod
Step 3
Build and deploy through one connected pipeline
Run builds, approvals, deployments, and rollbacks through repeatable workflows with the complete deployment history attached.
CopyDownload
Wrap
| 20:10:40.363 | info | Starting storefront server... | | ------------ | ---- | ------------------------------------------------------- | | 20:10:41.563 | info | Loading runtime configuration for production | | 20:10:42.763 | info | Initializing application routes and middleware | | 20:10:43.963 | info | Listening on: <http://localhost:4000/> (all interfaces) | | 20:10:45.163 | info | Application startup completed | | 20:10:46.363 | info | Health check passed: GET /health 200 |
Step 4
See what happened. Recover without losing context
Bring infrastructure changes, deployments, failures, and your existing logs and metrics together. Diagnose and recover through the UI, CLI, or API.
We outgrew Heroku, but direct AWS was unrealistic. I love Ravion’s agent-friendly structure for getting great defaults on AWS while still being able to customize deployments just the way we need. And their support has been invaluable.
Edgars Jekabsons, Engineering Manager / Ember
{From static sites to multi-cloud infrastructure}
Deploy web servers
Ship containerized web services to ECS with rolling, blue-green, canary, or linear rollouts, or update EC2 hosts in place over SSM. Add shared load balancers, TLS certificates, and pre- or post-deploy hooks.
Use it with
and many more
Deploy to KubernetesComing soon
Deploy background workers and AI agents
Deploy Lambda functions
Deploy static sites
Provision databases and caches
Provision networking, DNS, and certificates
Deploy infrastructure across any cloud
Loved by DevOps and Developers alike
We switched from Railway for better reliability and avoiding AWS console. But Ravion’s expert support is what sold me. I’m always surprised by the speed and quality of responses. It’s a no brainer for startups that want to move fast.
Yaser Alkayale, Co-founder / CivCore
Frequently asked questions
How is Ravion different from Terraform Cloud or Spacelift?
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
How is Ravion different from a PaaS?
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
Do I have to write Terraform?
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
Can I use CloudFormation, Pulumi, or CDK?
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
Which clouds does Ravion support?
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
Is support for GCP and Azure coming?
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.
Is there vendor lock-in?
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
What happens if Ravion goes away?
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
What happens if Ravion has downtime?
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
What access does Ravion need to my AWS account?
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
Can I bring infrastructure I already have?
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
Is Ravion safe for AI agents to use?
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
How old is the company?
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
What does Ravion cost?
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
Provision and deploy
your first app in minutesConnect your AWS account, pick a module, and ship.
Free to start / No credit card / Revoke access anytime