Question 1

What is CloudThinker?

Accepted Answer

CloudThinker builds Frontier Investigation Agents. They find the why behind every production issue and propose the fix. Your engineers approve. The agents investigate incidents, review code, manage infrastructure and cut cloud spend under policy your team sets. They work across AWS, Azure, GCP and Kubernetes with brokered credentials, sandboxed execution and a tamper-evident audit trail for every action.

Question 2

How are Frontier Investigation Agents different from AIOps?

Accepted Answer

Traditional AIOps platforms stop at alert correlation and anomaly detection. They surface insights for humans to act on. Frontier Investigation Agents go further: they investigate the root cause, propose the fix (a runbook step or a merge request) and check the outcome. Graduated autonomy lets your team decide how much each agent may do on its own in each environment.

Question 3

How is CloudThinker different from AI coding tools like Claude Code, OpenAI Codex, Amazon Kiro, or Cursor?

Accepted Answer

Coding tools generate diffs. CloudThinker is the operations platform that takes changes to production safely — answering the six questions a coding tool was never built to answer: who is connecting, what they can reach, how traffic gets there, what data crosses the boundary, what they did, and who said yes. Many teams use both: a coding tool to write the change, CloudThinker to operate the systems it lands on.

Question 4

Which AI agents does CloudThinker provide?

Accepted Answer

Specialist agents that share context. An orchestrator routes work and runs parallel investigations. Specialists cover cloud engineering (cost, configuration and IaC across AWS, Azure, GCP), security and compliance, Kubernetes (EKS, GKE, AKS) and databases (PostgreSQL, MySQL, Redis). You can also define custom agents and skills for your own workflows.

Question 5

How do I interact with the agents?

Accepted Answer

In plain language, from the CloudThinker web app, Slack or Microsoft Teams. Mention an agent (for example, ask it to find EC2 instances with low utilization) and it replies with analysis, dashboards, reports or a proposed action for approval. No new query language to learn.

Question 6

How quickly can we see results?

Accepted Answer

Setup takes about 10 minutes: connect a cloud account read-only and agents begin discovery immediately. Most teams see their first findings and recommendations the same day, measurable results within 48 hours, and material cost and toil reductions within the first month as they graduate agents from notify-only to approved actions.

Question 7

What is the Resolve?

Accepted Answer

Resolve is CloudThinker's Frontier Investigation Agent for incidents. It has three parts. It clusters raw alerts from tools like Datadog, Prometheus, Grafana and PagerDuty into deduplicated incidents. It runs parallel root-cause investigations. And Memory stores every resolution, so similar incidents get resolved faster over time.

Question 8

What is Optimize?

Accepted Answer

Optimize is CloudThinker's Frontier Investigation Agent for cloud cost. It sits above traditional FinOps dashboards. It runs a daily loop across AWS, GCP, Azure and Kubernetes: find waste, check it against live workload context, and ship the fix as a merge request with the diff, rationale, projected savings and rollback plan. It treats AI and LLM token spend as first-class, next to compute and storage.

Question 9

How does Review work?

Accepted Answer

Review checks every pull request in GitHub or GitLab for security vulnerabilities (OWASP Top 10), logic errors, performance issues and your coding standards. It posts inline comments with severity ratings and suggested fixes. It shares context with Resolve and the infrastructure agents, so reviews are incident-aware: it knows what has broken in production before.

Question 10

What is the AI Well-Architected Assessment?

Accepted Answer

An on-demand assessment that grades your AWS, GCP or Azure environment across the six Well-Architected pillars in about 10 minutes, using 135+ parallel AI agents reading live state. It is read-only and safe to run on production accounts. Its findings feed straight into Optimize and security remediation.

Question 11

What is the Skills Framework?

Accepted Answer

Skills are modular, composable capabilities agents use to do work: public skills maintained by CloudThinker, connection skills for each integration, and custom workspace skills that encode your own runbooks, standards, and tribal knowledge — created in natural language. The platform ships with 325+ pre-built runbooks covering common failure modes.

Question 12

What is Auto Mode, and what are the autonomy levels?

Accepted Answer

Auto Mode is the approval-gate layer. Each skill runs at one of four levels — Notify, Suggest, Act with Approval, or Autonomous — configured per environment and per service. The policy is encoded outside the agent, so an agent can never escalate its own permissions; high-risk actions always pause for a human, and every decision is audited.

Question 13

Which cloud platforms does CloudThinker support?

Accepted Answer

AWS, Microsoft Azure, and Google Cloud — including multi-cloud environments managed simultaneously — plus Kubernetes on EKS, AKS, and GKE. Multi-account and multi-region setups are supported out of the box.

Question 14

What tools does CloudThinker integrate with?

Accepted Answer

Over 100 integrations across the stack: source control (GitHub, GitLab, Azure DevOps, Bitbucket), chat (Slack, Microsoft Teams), on-call and ITSM (PagerDuty, Opsgenie, ServiceNow, Jira, Linear), observability (Datadog, Prometheus, Grafana, Splunk, New Relic, Dynatrace, Zabbix), IaC and delivery (Terraform, Pulumi, ArgoCD, Jenkins), and databases (PostgreSQL, MySQL, Redis, MongoDB, Kafka, Elasticsearch). Custom systems connect via the open MCP standard.

Question 15

Do we need to install agents in our environment?

Accepted Answer

No. CloudThinker connects through standard provider APIs and the MCP protocol — no daemons, sidecars, or agents to deploy. The recommended starter policy is read-only: agents can observe and recommend from day one, and write access is granted separately, gated by the approval workflow.

Question 16

How are credentials and connections managed?

Accepted Answer

Through Connections, the brokered-identity layer. CloudThinker never hands agents long-lived production credentials: a short-lived, scoped token is issued at task time, used inside an isolated sandbox, and revoked when the task completes. Credentials are encrypted at rest, follow least privilege, and every access is written to the audit log.

Question 17

Do you train AI models on our data?

Accepted Answer

No — never. CloudThinker uses enterprise-grade LLM endpoints (Anthropic, OpenAI Enterprise, AWS Bedrock, Azure OpenAI) under zero-retention contracts that contractually prohibit training. Your prompts, code, and operational data are used only to serve your requests.

Question 18

What compliance certifications does CloudThinker have?

Accepted Answer

CloudThinker is SOC 2 Type II certified with annual independent audits. The SOC 2 report, DPA, and security questionnaire are available from our trust center. HIPAA, ISO 27001, and PCI DSS readiness are supported through BYOC deployment, and the deterministic tokenization model addresses GDPR, CCPA, and regional regulations such as Vietnam Decree 13 and MAS Notice 658.

Question 19

How is sensitive data protected?

Accepted Answer

The Guardrails Engine redacts PII, PHI, and secrets from every prompt before it reaches a model provider, and deterministic tokenization replaces sensitive values with stable placeholders at the egress boundary — the LLM provider never sees the real value. Data is encrypted with TLS 1.3 in transit and AES-256 at rest, with per-workspace keys.

Question 20

What stops an agent from doing something destructive?

Accepted Answer

Layered safeguards: per-agent tool allowlists restrict what each agent may call; graduated autonomy requires human approval for high-impact operations by default; every action runs in an ephemeral, isolated sandbox with egress controls; and every decision is recorded with full reasoning so it can be reviewed or rolled back.

Question 21

Does CloudThinker support enterprise SSO and access controls?

Accepted Answer

Yes. SAML/OIDC single sign-on, SCIM user provisioning, enforced MFA, granular RBAC with custom roles, IP allowlists, and per-workspace isolation. Audit logs are immutable, tamper-evident, and exportable to your SIEM (Splunk, Datadog, Elastic, and others).

Question 22

What deployment options does CloudThinker offer?

Accepted Answer

Three. SaaS with private connectivity: the fully managed platform reaches your environment only over Site-to-Site VPN, VPC Endpoint (AWS PrivateLink), or VPC Peering — never the public internet. BYOC: the entire control plane and data plane run inside your AWS, Azure, or GCP account (or on-premise), fully managed by CloudThinker, so data never leaves your perimeter. Custom model: pin inference to your own LLM endpoints so prompts never leave your perimeter either.

Question 23

How does the SaaS platform connect to our environment?

Accepted Answer

Over private, mutually authenticated links — Site-to-Site VPN, VPC Endpoint (AWS PrivateLink), or VPC Peering — with no inbound public traffic required on your side. Combined with short-lived, scoped credentials per task, your production perimeter stays closed to the outside world.

Question 24

Can we use our own LLM or a custom model?

Accepted Answer

Yes. You can pin the platform to a single provider, bring your own keys (for example AWS Bedrock under your own billing), or point inference at a self-hosted or fine-tuned custom model — keeping all inference inside your security perimeter. Available on Enterprise plans and BYOC deployments.

Question 25

How does CloudThinker pricing work?

Accepted Answer

Per-seat plans with a monthly credit pool: Team at $25/seat/month, Scale at $100, Scale+ at $200 (each about 20% cheaper billed annually), and Enterprise with custom pricing for self-host/BYOC, customer-managed credentials, and your own LLM provider. Every plan includes the same agents and full product surface — higher tiers add credits, workspaces, SSO/SCIM, audit logging, and SLAs.

Question 26

What are credits, and what happens if we run out?

Accepted Answer

Credits meter agent work and are pooled across your whole team and workspaces. Team includes 200 credits per seat per month, Scale 1,000, and Scale+ 2,000. Overage is billed at $0.10 per credit on self-serve plans; Enterprise uses metered pay-as-you-go at a negotiated rate.

Question 27

Is there a free trial?

Accepted Answer

Yes — the Team plan includes a 14-day free trial you can cancel anytime. Enterprise evaluations typically run as a guided proof-of-concept with our team.

Question 28

What counts as a workspace?

Accepted Answer

One logical environment with its own agent configuration, rule set, and access controls. Teams typically split workspaces by environment (dev/staging/prod), business unit, cloud account, or region. Namespaces or microservices inside one environment do not count separately.

Question 29

How do we get started?

Accepted Answer

Create an account, connect your first cloud account from Settings → Connections (an IAM role with read-only access is recommended), and agents begin discovering your infrastructure within minutes. From there, ask your first question in chat — or book a demo and our team will walk your use case end to end.

Question 30

What support does CloudThinker offer?

Accepted Answer

Community support on the free tier, email support on Team, priority support with a 99.9% uptime SLA on Scale and Scale+, and dedicated support with a 99.95% SLA on Enterprise. Reach us anytime at support@cloudthinker.io.

Question 31

Where can I learn more?

Accepted Answer

The documentation at docs.cloudthinker.io covers setup, connections, and every module in depth. On this site, the Glossary defines the core concepts, the Skills Library shows what agents can do, and the blog covers real-world use cases — or talk to us directly.

CloudThinker FAQs

General

Product & Capabilities

Connections & Integrations

Security & Compliance

Deployment

Pricing & Billing

Getting Started & Support

Still have questions?