> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.qa.army/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.qa.army/_mcp/server.

# QA.army Chatbot

> Understand how conversational access fits QA.army and which boundaries keep product state trustworthy.

The QA.army chatbot is intended to make coverage and results easier to explore
in conversation while keeping authoritative product state in trusted QA.army
screens.

> **Warning**
>
> The SaaS chatbot is not currently available.

## Who it is for

| Reader | Useful question                                                               |
| ------ | ----------------------------------------------------------------------------- |
| CEO    | Which critical customer journey needs attention before launch?                |
| CMO    | Did the signup or checkout journey change after the campaign release?         |
| CTO    | Which failed assertion has reproducible technical evidence?                   |
| CPO    | Which high-value journey is missing durable coverage?                         |
| CIO    | Which Workspace members and connected workflows can access this Project?      |
| COO    | What failed, who owns the next action, and is the problem reproducible?       |
| CFO    | Which failures affect revenue-sensitive journeys, and what usage is in scope? |

## What conversation must not replace

The chatbot must not become a second source of truth. When it becomes
available:

* Workspace permissions will still control every request.
* The saved Test will remain the durable specification.
* The Run page will remain authoritative for status and evidence.
* A proposed action will use the same product operation as the SaaS app, CLI,
  or MCP.
* A conversational summary will not turn an `ERROR` into a product `FAILED`
  result.

## Target scope

Questions may concern Web, Android, iOS, desktop, or a journey that crosses
applications. Mentioning a target in chat will not make an unavailable
execution capability available. Mobile device execution, desktop execution,
and the chatbot itself are not generally available. Direct Web Runs can return
screenshots, recordings, and other captured evidence; inspect the actual Run
result and the evidence available for that attempt.

## Until the chatbot is available

Use saved Tests and Test Groups to record what matters. Use the SaaS app for
visual organization and Web Run review, the [CLI](/interfaces/cli) for terminal
workflows, or the published [local MCP server](/interfaces/mcp) with a supported
client. These interfaces use your existing Workspace access; the unavailable
chatbot is not needed to save a Test or inspect a Run.