# Comparison

## When to Use Stakpak vs ChatGPT, Gemini, Claude, and Copilot?

Stakpak.dev is an AI DevOps agent built to run *inside your infrastructure*. It connects directly to your cloud environment, infrastructure-as-code (IaC), and CI/CD systems, which allows it to understand, execute, and automate real operational tasks in a secure way

General LLM chat apps like ChatGPT, Gemini, Claude, and Copilot are conversational models. They are good at generating text, explaining concepts, and writing example code, but they don’t have direct access to your environment, can’t run commands, and don’t enforce safety controls (Like secret redaction)

TLDR:

* ChatGPT/Claude = Great for thinking and explaining.
* Stakpak = Great for executing and getting things done fast

### When to Use Stakpak for Infrastructure and DevOps?

**When:**

* Tasks that require understanding the context of your infrastructure
* You want actions, not suggestions
* Past deployments, policies, or incidents matter
* You need up to date docs
* You want to run it in the terminal or Slack
* Guardrails and safety are important

### When to use Chat GPT, Gemini, Claude, and Copilot for Infrastructure and DevOps?

* Ask questions that don't depend on the latest docs
* Learning DevOps/Infra Concepts
* Explaining technical stuff to the business team
* Planning & Research

## Stakpak vs Coding Agents

There are a lot of differences between coding tasks and operations tasks

|                                          | **Coding Tasks** | **DevOps Tasks**                 |
| ---------------------------------------- | ---------------- | -------------------------------- |
| **Feedback loop**                        | fast             | slow                             |
| **Users’ knowledge**                     | high             | < 3% of devs do this full-time\* |
| **Syntax/Language**                      | few popular      | many domain specific             |
| **Frequency of schema/knowledge change** | low              | high                             |
| **Context**                              | mostly local     | mostly remote                    |

<p align="right">*Stack Overflow Developer Survey 2024</p>

That's why stakpak is different from coding agents

* **Designed for Ops Tasks**: Handles infra tasks that coding agents struggle with.
* **Global playbooks**: [rulebooks](https://stakpak.gitbook.io/docs/how-it-works/rulebooks "mention") lets stakpak follow *your* own *standards*, not just generic best practices.
* **Remembers what matters**: [memory](https://stakpak.gitbook.io/docs/how-it-works/memory "mention") remember what worked so you don’t repeat yourself, and share these memories with your team.
* **Safe in production**: [warden-guardrails](https://stakpak.gitbook.io/docs/how-it-works/warden-guardrails "mention") prevent the agent from making destructive actions 100% of the time.
* **Async**: Supports long running, async operations without getting stuck.
* **Secure by design**: Secrets are substituted and never exposed to the LLM.
* **Transparent**: [agent-sessions](https://stakpak.gitbook.io/docs/how-it-works/agent-sessions "mention")full audit trail.