> For the complete documentation index, see [llms.txt](https://driftserver.gitbook.io/drift-server/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://driftserver.gitbook.io/drift-server/master.md).

# What is Drift Server

Have you as a software developer ever been in the following situation: you start working on a legacy code base. You have no clue what the application is doing or how it is implemented. Documentation is lacking or not up to date.  

Drift Server is a tool to help you in this kind of situation.  It is used to quickly gain some insights when you start working on a complex existing legacy code base. You can treat the whole system as a black box, start with a local database with as little data as possible and see what happens in the DB every time you do something in the front-end.

Say you want to know what the system is doing when you create a new user: 

1. Connect Drift Server to the same databases (or file systems) as  your legacy application.  

![The Drift  System Connections screen with an overview of the DB connections and file system directories](/files/-LlpcYPx9SP4j8xFaHmS)

2\. Start a new recording from the recordings page

![Recordings page](/files/-LkxHNc0w0jmMjMgvQqq)

3\.  Take an initial snapshot of the databases and filesystem as a baseline

![initial database database snapshot](/files/-LkxHOt_Zok7mO4QFMd3)

![initial filesystem snapshot](/files/-LlpfL7a9-gmlGXM9GSm)

4\.  Change the state of the legacy system by interacting with it (e.g. by executing the usecase "creating a new user") 

When done, take another snapshot, to compare with the first one to see what the application actually did:

![Database Delta showing the delta between two Database Snapshots](/files/-LkxHPtLvCpK7gxGmqOR)

![Filesystem Delta showing the delta between two  Filesystem snapshots ](/files/-Llqn9Yy5-UeF7kb-QXN)

Drift server is limited to a local development environment, with only a single user and with very low amount of data. It should work fine for about 10000 rows of data.