Skip to content
Kindo Kindo Docs

Frequently Asked Questions

What is Kindo?

Kindo is an AI-native automation platform for enterprise technical operations, including SecOps, DevOps, and ITOps teams. It uses intelligent AI agents and a domain-tuned LLM to automate operational workflows end-to-end. Kindo is model-agnostic and compatible with 26+ LLMs. Everything available in the UI is also accessible through a public API.

How is Kindo different from chatbot platforms?

Kindo is a full agentic automation platform. Unlike chatbot tools that only provide an AI assistant for answering questions, Kindo agents autonomously make decisions and take actions in your systems. They orchestrate multi-step workflows across your tech stack with appropriate approvals, rather than simply alerting a human.

What is an AI-native agent?

An AI-native agent in Kindo is an intelligent software assistant that understands natural language, reasons through tasks, and performs actions on your behalf. These agents encapsulate operational runbooks and playbooks. They can read human-written instructions and translate intent into programmatic actions like API calls, dynamically adapting to results at each step.

What is DeepHat?

DeepHat is Kindo’s proprietary LLM tailored for DevSecOps, SecOps, and infrastructure tasks. Trained on real-world IT incidents, cybersecurity scenarios, code logs, and CLI syntax, it excels at reading configurations, demonstrating exploits, and explaining root causes. It is designed for adversary simulation and red teaming.

Can Kindo run without internet access?

Yes. Kindo supports fully on-premises deployment with no internet connection required. All data, models, and operations stay behind your firewall. AI models like DeepHat can run locally on your GPU infrastructure.

What is MCP and how does Kindo use it?

The Model Context Protocol (MCP) is an open standard that allows AI agents to connect with external tools using a uniform JSON message format. Kindo uses MCP to integrate with virtually any system — source control, cloud services, ticketing systems, SIEMs, databases, and more — through a consistent interface.

How does Kindo secure MCP connections?

Kindo acts as a governance layer between AI and your tools:

  • Data Loss Prevention (DLP) — Scans all data passing through for sensitive information and automatically redacts disallowed content.
  • No Credential Exposure — The LLM never sees raw credentials. Secrets are injected at execution time without revealing them to the model.
  • RBAC — Administrators define exactly which actions agents can perform on which systems.
  • Whitelisted Integrations — Only approved, trusted MCP connections are allowed.
  • Audit Logging — Every agent action and tool access is logged for review.

What tasks can Kindo agents perform?

  • Vulnerability Management — Scan for CVEs, apply patches, trigger container rebuilds
  • IAM Enforcement — Audit roles and permissions, flag or auto-correct overly broad access
  • Incident Response — Triage alerts, gather logs, correlate root causes, contain issues
  • Infrastructure Automation — Rotate certificates, manage configurations, execute playbooks
  • Ticketing and Orchestration — Escalate tickets, create issues, orchestrate cross-tool workflows
  • Compliance Auditing — Check configurations against benchmarks, gather audit evidence

Do I need to rewrite my runbooks?

No. Kindo augments your existing runbooks and playbooks. You can feed human-written procedures to an agent as instructions, and it will follow them. Kindo agents can also execute existing scripts and tools via integrations.

How does human-in-the-loop work?

You can configure approval checkpoints where agents must pause and request human sign-off before taking certain actions. Operators review proposed actions and approve or deny with a single click. Escalation paths and full audit trails are built in.

What models can I run?

Kindo is model-agnostic. It includes DeepHat and supports models from OpenAI (GPT-4), Anthropic (Claude), Google (Gemini), Meta (Llama), and more. You can use different models for different tasks within the same platform. All models operate under Kindo’s security framework with DLP and audit logging.

Can I run Kindo on premises?

Yes. Kindo offers fully on-premises deployment via Kubernetes Helm charts. You maintain full control over all data, models, and operations. The on-prem version includes all features — the agent framework, DeepHat, integrations, and everything else. Terraform plans are included for major cloud providers.

What security measures are built in?

  • RBAC — Role-based access for users and agents
  • DLP — Real-time monitoring and redaction of sensitive data
  • Audit Logging — Comprehensive trails of all actions and decisions
  • Encryption — Data encrypted at rest and in transit
  • Policy Engine — Configurable guardrails and action restrictions
  • Secrets Management — Secure credential storage with need-to-know access

Does Kindo support compliance frameworks?

Kindo is SOC 2 Type 2 compliant and supports frameworks like ISO/IEC 27001, GDPR, and NIST CSF through its built-in security features. For FedRAMP requirements, Kindo can be deployed in approved cloud environments like AWS GovCloud under your own controls.

How does Kindo integrate with my tools?

Kindo uses MCP and native integrations to connect with source repositories (GitHub, GitLab), CI/CD pipelines, cloud platforms (AWS, Azure, GCP), ITSM tools (Jira, ServiceNow), monitoring systems (Datadog, Splunk), and many more. Integrations are bi-directional and can be chained within a single workflow.

What is a secure sandbox?

A secure sandbox is an isolated VM created for every agent conversation. Agents use sandboxes to run shell commands, execute Python scripts, download and transform data, and create documents. Sandboxes are hardened, never contain credentials, and are destroyed after inactivity.