Product

CloudPilot

Intelligence, visibility and control for your cloud.

A multi-cloud management platform that brings FinOps, SecOps, Compliance and infrastructure as code (IaC) together, with artificial intelligence to support analysis, recommendations and more efficient decisions.

CloudPilot panel with the cost, security and compliance dashboard.
AI
Amazon Web Services
Microsoft Azure
Google Cloud
CloudPilot interface. Screenshot from the demonstration environment.

Challenges in the cloud environment

Three challenges that cost a lot in a multi-cloud environment.

Environments spread across different clouds bring complexity, risk and unnecessary costs. CloudPilot exists to simplify that scenario.

Uncontrolled costs

Hard to spot spending and real optimization opportunities: invoices with hundreds of lines, inconsistent tags and undefined cost allocation per team.

Security risks

Exposure to threats, misconfigured access and lack of monitoring in every environment, with no single view of risk.

Complex compliance

Hard to meet standards and policies, collect evidence and keep audits current across multi-cloud environments.

How CloudPilot helps

Four steps. One conversational panel.

From connecting accounts to running actions, every step preserves traceability and offers explicit human approval points.

  1. Ask

    Ask questions in natural language about your data.

    How it works inside
    Connect AWS accounts (cross-account assume-role, SSO or keys), GCP (service account) and Azure. Credentials are encrypted at rest (Fernet + KMS) and permissions are read-only by default.
  2. Analyze

    CloudPilot cross-checks data and identifies trends and anomalies.

    How it works inside
    Scheduled jobs ingest costs, inventory, findings and metrics across the three clouds into a single model: costs in the FOCUS format (AWS CUR/S3, GCP BigQuery Billing Export, Azure Cost Management). The AI narrates; without evidence, the agent refuses to infer a root cause.
  3. Get insights

    Clear, prioritized recommendations, with evidence.

    How it works inside
    Every finding carries severity, evidence and a copy-ready remediation command. Multi-provider LLM (Gemini, Anthropic, OpenAI, Groq, Mistral, DeepSeek, Qwen, MiniMax) with streaming chat (SSE) and an autonomous Pilot Agent (WebSocket).
  4. Act

    Track results and automate routines with governance.

    How it works inside
    Opportunities become actions with human approval, dry-run, snapshot and rollback as first-class features. Approval via Slack or Teams: the approver cannot be the requester. Every execution leaves an audit trail.

Practical example

See how a question becomes a grounded answer.

CloudPilot combines your data, standards and cloud provider information with AI to answer clearly, point to evidence and suggest actions.

Diagram of the eight CloudPilot modules: FinOps, SecOps, Compliance, Drift, IaC, Governance, Reports and AI Agent, connected to the CloudPilot core.

Integrated modules

Eight modules covering the full multi-cloud cycle.

CloudPilot brings FinOps, SecOps, Compliance and IaC together with artificial intelligence, in modules that deliver actions and results in a single environment.

Multi-cloud coverage

One standard, for every environment.

CloudPilot connects to the leading cloud providers, offering a unified view regardless of the environment or the number of accounts.

Supported
  • Costs in the FOCUS format from CUR and S3
  • Continuous security scanning and IaC drift
  • Resources in a single panel, with budgets and chargeback
Supported
  • Costs in the FOCUS format via Azure Cost Management
  • Unified policies, security and compliance
  • Analysis and recommendations in the same panel
Supported
  • Costs in the FOCUS format via BigQuery Billing Export
  • CO₂e estimate per region (GreenOps), derived from ingested cost
  • Standardized reports and workload cost allocation

AWS, GCP and Azure are treated as supported. Capability coverage is validated continuously — the matrix below shows the validation status of each item.

Capability coverage matrix (validation status)
Capability AWS GCP Azure
Costs / FinOps (FOCUS)Under validationUnder validationUnder validation
Security (continuous scanning)Under validationUnder validationUnder validation
Compliance (CIS, SOC 2, PCI)Under validationUnder validationUnder validation
Drift & IaCUnder validationUnder validationUnder validation
Kubernetes FinOpsUnder validationUnder validationUnder validation
GreenOpsUnder validationUnder validationUnder validation
Automation (OpenOps)Under validationUnder validationUnder validation

CloudPilot is currently hosted on infrastructure separate from AWS. The adoption of AWS services for operational evolution is part of the technical plan.

Three-dimensional shield and servers on a dark blue background, with four icons: protected and encrypted data, access control and permissions, compliance with standards and policies, audit and traceability.

Security and governance

Built with the same standard we apply to our clients.

Controls already operating in the product

CloudPilot follows security, privacy and governance practices, focused on protecting your clients' data and meeting market standards.

  • Row-Level Security in PostgreSQL per workspace
  • JWT with refresh rotation
  • Fernet credentials + Infisical KMS
  • Fail-closed rate limiting
  • Audit log
  • Approval + owner/admin RBAC
  • Inputs validated with Pydantic v2
  • Anti-enumeration password reset
  • Linearized assume-role revocation (fail-closed)
  • Two kill switches · fail-closed

The scale behind reliable results

Results that build confidence.

Numbers measured directly in the product code. The test counts are functions and test blocks, not cases collected by the suite. These are product development metrics — not customer results.

419
REST endpoints
57
API modules
140
migrations
57
scheduled jobs
6.568
backend tests
1.515
frontend tests
18
E2E specs
8
LLM providers
Stack: Python/FastAPI · React 19/TypeScript · PostgreSQL 16 + pgvector · Redis · Celery · OpenTelemetry

A real gain for your team

Continuous evolution, with early access.

CloudPilot is under continuous development, with new features shipped every cycle. No commercial availability date has been announced.

Early access

Be the first

To test new features and shape the roadmap, alongside the people building the product.

More intelligence

For your operations: more context, more recommendations and more results in less time.

One platform

That evolves with your company's needs, without rebuilding the integration on every change.

Contact

Shall we talk about CloudPilot?

Early access is a conversation. Send your questions, your use case, or request a demo.

Where does CloudPilot run?

It runs on infrastructure operated by E F Cunha Soft House, currently separate from AWS. The adoption of AWS services for hosting and evolution is part of the product's technical plan.

Which clouds are supported?

AWS, GCP and Azure are treated as supported. The capability coverage matrix above marks the cells that are still pending validation.

Does the AI change my infrastructure without me asking?

No. The AI recommends and only executes with explicit human approval (via Slack or Teams). Write actions are opt-in, disabled by default, and there are two fail-closed kill switches to stop any execution.

Where are my credentials stored?

Cloud credentials are encrypted at rest (Fernet + KMS via Infisical) and used for reading only by default. Write permissions require additional approval.

How do I request early access?

Fill in the form below or write directly to efcunha@edsoncunha.eti.br with your use case.

This site is static. The form opens your email client with the fields filled in. Nothing is stored here.