Through cloud APIs
Create and update resources directly through your cloud provider’s APIs. Use this route for resources you authorise Latifold to manage.
Your agent builds.
Your cloud keeps up.
As your coding agent adds a feature, Latifold creates or updates the infrastructure it needs in your existing cloud. The agent can act automatically within your permissions and policies, with approval only when your rules require it.
Two ways to change your cloud.
Your coding agent connects to Latifold through MCP or API. Latifold then applies infrastructure changes through cloud APIs or the resource’s existing infrastructure-as-code workflow.
Through infrastructure as code
Make changes through your existing Terraform or other IaC workflow, keeping its code as the source of truth.
Every change follows the same planning, policy checks, approval when required, verification and recorded history.
Agents make changes.
You set the boundaries.
You choose the accounts, resources and operations your agents can work with. Latifold checks ownership, the change plan and deterministic policies before execution.
Agents carry out permitted changes themselves. Human review is needed only when your rules require it, and approval applies to that exact plan. If the plan or relevant environment changes, Latifold revalidates it.
Explore policies and controls- Within your rules
- Runs automatically.
The agent applies the change through Latifold, then verifies the result.
- Approval required
- Waits for your review.
An authorised person approves the specific plan before execution.
- Policy denied
- Does not run.
The change must satisfy your policy. An agent cannot override a denial.
Estimated cost thresholds can trigger approval. Actual spending enforcement is a separate control.
Keep building in your editor.
Keep oversight in your dashboard.
You describe the feature. Your agent handles the infrastructure requirement through Latifold and brings the configuration back into the application.
In your coding agent
Stay in Codex, Claude Code or Cursor. Your agent connects through MCP or API to inspect existing resources, understand relationships and ownership, and make the changes its code needs.
Available logs, metrics and recent changes provide operational context. The agent receives the resulting configuration so it can connect and test the feature.
In your dashboard
Connect cloud accounts, set agent permissions and policies, review changes that need approval, and inspect execution and audit history.
Each change keeps its plan, policy decision, approval when needed, provider results and verification together. Partial failures and unresolved differences remain visible for the next task.
Infrastructure context and change history carry forward between agent sessions.
Use the cloud you already have.
AWS, Azure and Google Cloud are supported; available resources and operations vary by provider. Start with read-only discovery and grant write access for the work you want agents to handle. Start without IaC, or work with the infrastructure code you already use. No infrastructure rebuild required.
AWS
SupportedConnect your existing AWS account. Begin with read-only discovery and add narrowly scoped write access for supported operations.
Azure
SupportedConnect your existing Azure account through the same Latifold interface. Begin read-only and grant write access for supported resources and operations.
Google Cloud
SupportedConnect your existing Google Cloud account. Discover supported infrastructure and let agents make changes within your permissions and policies.
Keep each resource’s source of truth.
Latifold checks who owns a resource before changing it. Existing Terraform, OpenTofu, CloudFormation and CDK workflows remain authoritative.
Observed
Read-only discovery, diagnostics and recommendations. Latifold does not automatically correct these resources.
Externally managed
Terraform, OpenTofu, CloudFormation, CDK or another owner stays authoritative. Proposed changes go through its existing review and deployment workflow.
Latifold-managed
Latifold's versioned desired state is authoritative. Authorised changes use native cloud APIs and controlled reconciliation within the agreed scope.
Latifold’s resource model keeps context and desired configuration centrally versioned. JSON/YAML export is planned for portability.
Keep coding. Let your agent handle the infrastructure.
Give your coding agent context on existing infrastructure and the ability to make changes within your permissions and policies. Changes proceed automatically unless your rules require approval.