Mongo Playground

What it is

Mongo Playground describes itself as "a simple sandbox to test and share MongoDB queries". The page has two panes: on the left you define the data, on the right you write a query, and the result appears below. Its README says queries are executed on an Atlas cluster provided by MongoDB, so you are running against a real database rather than an imitation.

It is an independent open source project, inspired by the Go Playground, and the source is on GitHub under the AGPL-3.0 licence.

What you can do in it

• Templates: start from a single collection, multiple collections (useful for trying $lookup), an update, indexes or an explain template.
• Data: type documents as BSON-style JSON, or generate them with the mgodatagen mode when you need more varied data.
• Queries: find(), aggregate() and update() are supported.
• Share: save a playground and get a link. This is why you see Mongo Playground links all over Stack Overflow answers about aggregation.

Why it is worth using

The fastest way to understand an aggregation stage is to change it and watch the output change. Here you can do that in seconds, on exactly the handful of documents you care about, with nothing to install. When a pipeline behaves oddly in your application, cutting it down to a small playground usually reveals the problem, and if it does not, the link is a perfect minimal example to share when you ask for help.

Limits to know

• A database can hold at most 10 collections and a collection at most 100 documents, so it is for small experiments, not performance testing.
• Only find, aggregate and update run. Inserts, deletes and admin commands do not.
• Shell regex literals such as /pattern/ do not work; use { $regex: "pattern" } instead.
• The page shows which MongoDB version it runs as "unknown" at times, so a very new operator may not be available.
• Shared playgrounds are saved forever and anyone with the link can open them. Never paste real customer data, keys or passwords.

Common questions

Is it a replacement for mongosh or Atlas? No. It is a scratchpad. Use it to test an idea, then run it against your own database.

Is it official? No, it is a community project, though it runs on infrastructure MongoDB provides.

Best for: learners practising queries and aggregation, and anyone who needs a small reproducible example.