---
title: "FAQ, MCP & limits"
section: "Analyze"
description: "Answers to common questions about SmartestQA Analyze, how AI assistants write requirements through MCP, and the module’s limits."
url: https://docs.smartestqa.ai/en/analyze-faq.html
audience: [customer, admin]
updated: 2026-11-08
keywords: "analyze faq questions requirement sub requirement ai data jira confluence quality score mcp smartest_propose smartest_apply limits"
---
# Analyze: FAQ, MCP & limits

Answers to common questions about Analyze, how AI assistants use this module, and its limits.

## In short

**On this page you can**

- Get quick answers about requirements, AI, links, roles and data.
- Filter the questions for customers or admins.
- Let an AI assistant write requirements through MCP, with your approval.
- Look up the limits that apply to imports, AI and requirements.

**How to use it**

1. Pick your audience with the filter above the questions.
2. Open a question to read the answer and its links.
3. For assistants, connect one from Device Farm › Test Agents › Connect to Claude.
4. Check the limits table before importing a large document.

## Questions and answers

### Q: What is the difference between a main requirement and a sub-requirement?

A main requirement is a feature or user flow and sits in the tree. A sub-requirement is one condition with an expected result inside it, and that is where test cases are linked. See [building blocks](https://docs.smartestqa.ai/en/analyze.html#concepts) and [a good example](https://docs.smartestqa.ai/en/analyze-requirements.html#write-good).

_Audience: customer_

### Q: Why can’t I link a test case to a main requirement?

Links belong to sub-requirements, so every test proves one specific condition. The main requirement shows the test cases of all its subs together. Select a sub to add one: [Link a test case](https://docs.smartestqa.ai/en/analyze-test-cases.html#link-existing).

_Audience: customer_

### Q: Can I use Analyze without AI?

Yes. Folders, requirements, sub-requirements, links, documents and comments work without AI. The AI wizard, **Generate**, **Automate** on a sub and [**Assessment (AI)**](https://docs.smartestqa.ai/en/analyze-assessment.html#run) need [**Enable AI**](https://docs.smartestqa.ai/en/project-settings-ai.html#enable-ai) (and the wizard also [**AI Document Analysis**](https://docs.smartestqa.ai/en/project-settings-ai.html#what-each-switch-unlocks)) in [Project Panel › AI Options](https://docs.smartestqa.ai/en/project-settings-ai.html#capabilities).

_Audience: customer, admin_

### Q: Which files can I import, and how big can they be?

Word, PDF with text, text, Markdown, HTML, JSON, CSV, YAML, XML, Excel and common image formats, several at once; see [Import](https://docs.smartestqa.ai/en/analyze-import.html#step-import). There is no fixed document size, but the browser waits at most 3 minutes for the outline, so split long documents by chapter or module.

_Audience: customer_

### Q: Does the AI read the documents in the Documents folder?

Not the in-app AI: it reads only the files you attach on the [Import tab](https://docs.smartestqa.ai/en/analyze-import.html#step-import). An AI assistant connected through [MCP](https://docs.smartestqa.ai/en/analyze-faq.html#mcp) can read and write those documents.

_Audience: customer_

### Q: Does a high Quality Score mean the requirement is tested?

No. The score rates how clear and complete the requirement text is. Whether it is tested shows in its linked test cases and in [InSight › Test Health](https://docs.smartestqa.ai/en/insight-test-health.html#pillars). See [Quality Score](https://docs.smartestqa.ai/en/analyze-numbers.html#quality-score) and [What counts as covered](https://docs.smartestqa.ai/en/analyze-test-cases.html#coverage).

_Audience: customer_

### Q: Does the Status change when the tests pass?

No. Status (New, Assigned, Implemented, Closed) is set only by people. Test results show in [Report](https://docs.smartestqa.ai/en/report.html) and [Test Health](https://docs.smartestqa.ai/en/insight-test-health.html).

_Audience: customer_

### Q: Does Analyze work for mobile and desktop projects?

Yes, the same way. Two parts of the AI wizard assume a web app: **Record Actions** uses the [browser extension](https://docs.smartestqa.ai/en/quick-start-connect.html#install-extension), and the optional pipeline needs a [Start Page URL](https://docs.smartestqa.ai/en/quick-start-connect.html#start-page-url). For mobile and desktop, capture missing controls in [Discover](https://docs.smartestqa.ai/en/discover-mobile-desktop.html) and [create pipelines in Schedule](https://docs.smartestqa.ai/en/schedule-create.html#create).

_Audience: customer_

### Q: Can I get back a sub-requirement I deleted?

No. A sub removed in the list and saved is deleted for good, with its links. Main requirements and folders go to the Archive and can be restored. See [Archive and restore](https://docs.smartestqa.ai/en/analyze-folders.html#archive).

_Audience: customer_

### Q: Can I import requirements from Jira or Confluence?

From Confluence, only through an AI assistant: the [Confluence page](https://docs.smartestqa.ai/en/integrations-confluence.html#import-requirements) shows **Requirements Import** as **Coming soon**, with no button. An assistant connected through [MCP](https://docs.smartestqa.ai/en/analyze-faq.html#mcp) can start the import, which creates main requirements in a **Confluence Import** folder. Analyze has no Jira import. You can export the issues to a file and use the [AI wizard](https://docs.smartestqa.ai/en/analyze-import.html), or let an AI assistant write them through [MCP](https://docs.smartestqa.ai/en/analyze-faq.html#mcp).

_Audience: customer, admin_

### Q: Do requirements count toward our licence?

Yes. Your company’s licence can limit how many requirements you create. Creating a requirement or a document from the tree adds one to the company’s requirement counter, and archiving does not take it back. At the limit you see “Company license limit has been reached…”. An Admin sees the licence in [Company Settings › Billing](https://docs.smartestqa.ai/en/company-settings-billing.html).

_Audience: admin, customer_

### Q: Who can delete requirements, folders and modules?

Deleting moves items to the Archive. Users can archive what they created entirely; [Admins](https://docs.smartestqa.ai/en/company-settings-roles.html#matrix) can archive anything, including modules, and only Admins can delete for good. See [Who can do what](https://docs.smartestqa.ai/en/analyze-folders.html#who-can).

_Audience: admin_

### Q: What does Analyze send to the AI provider?

The wizard sends the attached documents and images, the actions and [App Map descriptions](https://docs.smartestqa.ai/en/discover-app-map.html#page-information) of the scope (with up to 6 screenshots), template test cases, similar existing requirements and active [Notes](https://docs.smartestqa.ai/en/discover-notes.html#notes-to-ai). The assessment sends the names, descriptions and severities of a main requirement and its subs. Everything goes to the provider configured in [AI Options](https://docs.smartestqa.ai/en/project-settings-ai.html#data). Use test data in documents.

_Audience: admin_

### Q: Which AI Options does Analyze use?

[**Enable AI**](https://docs.smartestqa.ai/en/project-settings-ai.html#enable-ai) for every AI feature here, plus [**AI Document Analysis**](https://docs.smartestqa.ai/en/project-settings-ai.html#what-each-switch-unlocks) for the Import wizard. The [AI account](https://docs.smartestqa.ai/en/project-settings-ai.html#ai-account) (SmartestQA’s or your own key), the [provider and the model](https://docs.smartestqa.ai/en/project-settings-ai.html#own-account) come from the same settings.

_Audience: admin_

## For AI assistants (MCP)

An assistant [connected through the SmartestQA MCP server](https://docs.smartestqa.ai/en/mcp-connect.html) works on Analyze with [two calls](https://docs.smartestqa.ai/en/mcp-safe-changes.html#propose-apply): `smartest_propose` returns a diff and writes nothing; after you approve, `smartest_apply` writes exactly that preview. Connect an assistant from **Device Farm** › **Test Agents** › [**Connect to Claude**](https://docs.smartestqa.ai/en/mcp-connect.html#claude-desktop).

| Level | Parent | What the assistant can do |
| --- | --- | --- |
| `main` | A module | Create, enrich, update, remove main requirements; reorganise them between modules |
| `sub` | A main requirement | Create, enrich, update, remove sub-requirements, with their source and a status line (Verified, Inferred, Conflict, Not in code); mark a sub as not automatable with a reason |
| `document` | Documents | Write project documents; `smartest_discover` with `kind=documents` reads them |

- `smartest_context` with a main or sub as target [tells the assistant where things stand](https://docs.smartestqa.ai/en/mcp-tools.html#tools-overview); `smartest_status` shows what is missing project-wide.
- A sub-requirement marked not automatable is left out of [Test Health coverage](https://docs.smartestqa.ai/en/insight-numbers.html#pillars). The reason can be up to 1,000 characters; only subs can be marked.
- An assistant can delete a requirement only if it created it, and only when the connected user is an [Admin](https://docs.smartestqa.ai/en/company-settings-roles.html#matrix).

Example prompts

```text
Read the checkout code in /web/src/checkout and propose main requirements for the module Payments, with sub-requirements for every validation rule.

Which sub-requirements in Payments have no test case yet? Propose one scenario for each.

Mark the sub-requirement "Bank settlement file is received" as not automatable: it needs the bank's batch system.
```

## Limits

| Item | Limit |
| --- | --- |
| Requirement name | 255 characters |
| Sub-requirement description written by an MCP assistant | 255 characters |
| Not-automatable reason (MCP) | 1,000 characters |
| Browser wait: outline · step generation | 3 min · 10 min |
| Server budget for one step generation | 420 s, batches of 4 scenarios |
| App Map screenshots per analysis (with scope) | 6 |
| Similar requirements sent to avoid duplicates | 10 |
| Proposals from Generate Sub Requirements | 3–5 |
| Proposals from Automate on a sub | 1–3 test cases |
| Sub description for Automate | at least 10 characters |
