> ## Documentation Index
> Fetch the complete documentation index at: https://docs.monk.io/llms.txt
> Use this file to discover all available pages before exploring further.
# Cloud Infrastructure
> Autonomous cloud resource setup across providers
## What It Does
You describe what you need. Your coding agent and Monk figure out the infrastructure — VMs, networks, storage, load balancers, GPUs — and provision it across any cloud provider. No cloud consoles. No Terraform. No provider-specific tooling.
Monk reads your application's structure and requirements, then creates right-sized resources in the best regions for your workload.
## Supported Cloud Providers
Monk works with your existing cloud accounts:
* [AWS](https://aws.amazon.com/) — Full support
* [Google Cloud Platform](https://cloud.google.com/) — Full support
* [Microsoft Azure](https://azure.microsoft.com/) — Full support
* [DigitalOcean](https://www.digitalocean.com/) — Full support
* [Hetzner](https://www.hetzner.com/cloud) — Full support
**Need another provider?** [Let us know](/getting-help). Monk's architecture
makes it straightforward to add new cloud platforms.
## How It Works
### Connect Your Cloud Account
Monk uses your existing cloud accounts. No new accounts needed.
1. Monk detects it needs to provision resources
2. You're asked for cloud credentials for the target provider
3. You provide credentials once per provider
After that, Monk provisions and manages infrastructure automatically. No further cloud console work required.
**Your credentials stay secure.** Credentials are encrypted at rest in your IDE's secret storage and on your Monk cluster using your cloud provider's KMS. They are never sent to Monk servers and never exposed to the LLM. See [Security](/features/security) for full details.
> [Step-by-step credential guides](/getting-started/obtaining-credentials) for AWS, GCP, Azure, DigitalOcean, and Hetzner
### Intelligent Resource Selection
Based on [code analysis](/features/code-analysis), Monk understands what your app needs and provisions accordingly.
**Instance sizing:**
* **API servers** — sized for your expected traffic and compute needs
* **GPU workloads** — appropriate GPU instances for ML/AI services
* **Memory-intensive services** — sufficient RAM for databases, caches
* **Workers** — right-sized for background job processing
**Region selection:**
* **Performance** — minimizes latency to your users
* **Cost** — balances performance with budget
* **Compliance** — respects data residency requirements
* **Multi-region** — distributes across regions when needed
**You confirm before anything gets created.** Monk shows the proposed infrastructure. You can discuss, adjust, and approve — all in chat.
**Example conversation:**
```
Monk: Your API needs 2 VMs for the backend. I'm recommending:
- 2x n1-standard-2 (2 vCPU, 7.5GB RAM) in us-central1
- Estimated cost: ~$100/month
You: Can we use a cheaper region?
Monk: Sure! Using us-central1 (Iowa) is 15% cheaper than us-east1 (Virginia).
Updated estimate: ~$85/month. Sound good?
You: Perfect
Monk: Creating resources...
```
### Core Resources
Monk provisions and manages these automatically.
**Compute:**
* **Virtual machines** — right-sized instances for your workloads
* **GPUs** — for AI/ML, rendering, or compute-intensive tasks
**Storage:**
* **Persistent volumes** — for databases and stateful services
* **Snapshots** — automatic backups with point-in-time recovery
* **Block storage** — attached to instances as needed
**Networking:**
* **VPCs** — isolated networks for your application
* **Subnets** — proper network segmentation
* **Security groups and firewalls** — access control
* **Load balancers** — traffic distribution and high availability
### Beyond Core Resources
Monk handles specialized cloud services too. See [Integrations](/integrations) for the full list:
* Managed databases (RDS, Cloud SQL, Azure Database)
* Object storage (S3, Cloud Storage, Azure Blob)
* Message queues (SQS, Pub/Sub, Service Bus)
* Caching services (ElastiCache, Memorystore)
* DNS services (Route 53, Cloud DNS, Azure DNS)
When your application needs one of these, Monk integrates it automatically.
### Intelligent Placement
Monk co-locates resources with their dependencies.
* Redis Cloud in GCP `eu-central`? Monk places your containers in the same region.
* AWS RDS in `us-west-2`? API servers go to `us-west-2`.
* External API in Europe? Services placed nearby for low latency.
It also understands third-party service locations (Redis Cloud, MongoDB Atlas, etc.), minimizes inter-cloud data transfer costs, and optimizes for latency when crossing provider boundaries.
### Multi-Region and Multi-Cloud
You can use one region for simplicity, or span multiple regions for global reach. Monk handles the complexity either way.
Not sure which region? Ask. "Which region should I use for US customers?" Monk recommends based on latency, cost, and availability.
You can also run parts of your application on different providers simultaneously — frontend on one, API on another, database on a third. See [Multi-Cloud Support](/features/multi-cloud) for details.
### Continuous Infrastructure Management
Monk doesn't just provision once. It manages infrastructure throughout your application's lifecycle.
**While your app runs, Monk:**
* **Adds resources** — scales up when needed
* **Removes unused resources** — cleans up to save costs
* **Tracks costs** — real-time spending per resource. See [Cost Tracking](/features/cost-tracking).
* **Optimizes placement** — adjusts based on usage patterns
* **Maintains security** — updates firewalls and security groups as needed
**You stay in control.** Monk asks before making significant changes. You can tell it to scale up, scale down, or optimize. Nothing happens without your approval.
## Next Steps
Deploy across multiple providers simultaneously.
How your infrastructure and credentials are protected.