# What Apifolio does not do, and why

No live requests, no URL or Git sources, no stored credentials, no external files: what Apifolio leaves out to keep your API specs inside Confluence, and what to do instead.

Apifolio runs on Atlassian: it has no servers of its own and sends nothing outside Atlassian. That rule, and a short list of permissions, explain everything on this page. Each item says what to do instead.

## No "Try it out" (no live requests)

A live request from Confluence would send your data, and often your credentials, from your site to a server outside Atlassian. The API would also have to allow cross-origin calls from Confluence (CORS).

**Instead**: [Edit request](/docs/edit-request/) fills in the parameters, body, server and credentials, and gives you the request in [17 languages](/docs/code-samples/), ready to run in your terminal or API client. **Copy for AI** gives your assistant the same request.

## No specs loaded from a URL or a Git repository

Fetching a file from GitHub, GitLab, Bitbucket, SwaggerHub or any URL would be a call outside Atlassian, and it would need tokens stored somewhere.

**Instead**: attach the file to a page. Upload new versions from your CI pipeline, with your own token in your own CI: see [Keep the docs in sync](/docs/keep-in-sync/).

## No stored credentials

Apifolio never stores API keys, tokens or passwords, in the page or anywhere else. Code samples use placeholders such as `YOUR_TOKEN`. What a reader types in Edit request stays in that browser tab's memory and is gone when the editor closes.

**Instead**: readers use their own credentials in their own tools.

## No external `$ref` files or images

A spec split over several files (`$ref: ./schemas/order.yaml`) would need Apifolio to fetch the other files. References to other files and URLs are never followed: a schema shows "External reference … (not loaded)" instead, and a spec whose paths all live in other files lists those files with a tip on bundling. Images in descriptions and logos hosted on the web are shown as links, never loaded.

**Instead**: bundle the spec into one file before you attach it, for example with `npx @redocly/cli bundle openapi.yaml -o bundled.yaml`. A logo can be a `data:` image in the spec, or an image attached to the same page as the spec.

## No AsyncAPI

Apifolio shows OpenAPI 3.0, 3.1 and 3.2 and Swagger 2.0 documents. AsyncAPI documents are not supported.

## No editing of the spec

Apifolio shows the spec file; it doesn't change it. Your spec lives where you edit it: your repository, your editor or your API design tool.

**Instead**: upload a new version of the file (by hand or from CI). In the Apifolio dialog, **Paste** with the same file name also saves a new version of the attachment.

## No automatic changes to your pages

Apifolio asks for no permission to edit pages. So:

- the Confluence search index of a macro is refreshed when an editor saves the macro, not when the file changes. Editors see a note while editing the page when the index is older than the file. See [Confluence search](/docs/confluence-search/).
- importing another app's macro works one page at a time, in the Apifolio dialog, and the old macro is deleted by hand. See [Moving from another app](/docs/migration/).

## No content on Confluence public links

Confluence public links don't display Marketplace app content, Apifolio included.

**Instead**: choose ⋯ → **Copy as Confluence content** and paste the snapshot into the page, or export the page to PDF. See [Public spaces](/docs/public-and-anonymous/).

## No tracking

Apifolio contains no analytics, no tracking and no calls home. Your choice of code language is remembered in your own browser only.

## No data outside Atlassian

Your specs, pages and searches stay in Confluence. Apifolio functions run on Atlassian's infrastructure, and the Rovo agent runs on Atlassian's AI. See [Security](/security/) and [Privacy](/privacy/).
