# MongoDB Atlas cluster setup: a free cloud database > Published 2026-10-26T00:41:54.987Z on https://youblog.app/p/mongodb-atlas-cluster-setup-a-free-cloud-database/ > Source video: https://www.youtube.com/watch?v=upaF_Q1SXJQ MongoDB Atlas cluster setup changed less than you might expect. The console layout moved, the free tier gained and lost features, and the default connection method shifted from a raw connection string to the driver-first connect dialog, but the same three objects still have to exist: a cluster, a database user and an IP allowlist entry. ## MongoDB Atlas cluster setup in three minutes MongoDB Atlas cluster setup means creating a hosted MongoDB cluster, a database user, an IP allowlist entry and a connection string that your application uses to reach the database. The free tier is called M0 and needs no payment card. A 2020 video by the Brazilian channel Dev doido showed the older console flow; the object sequence is unchanged in 2026. MongoDB Atlas is MongoDB's managed database service, running on AWS, Google Cloud or Azure. You do not install a server, patch it or size a machine for a learning project. Atlas handles that, and the free cluster never expires, though it exposes only a subset of Atlas features. The video walkthrough is short and practical. It registers an account, creates a project, deploys the free cluster on AWS in a nearby region, then creates a user and a connection string. Those four actions are still the core of the process. What moved is where the console puts them. One detail from the video still holds in the documentation: each Atlas project can contain only one free cluster. If you want a second free database, create a second project. ## What the free M0 tier includes and where it stops The M0 free tier is a small shared development environment that never expires, and it gives access to only part of Atlas. Anything that requires dedicated hardware, full backup configuration or advanced operational controls needs a paid tier. Treat M0 as a sandbox for learning and prototypes. Atlas documentation separates the free cluster from paid clusters, which get full access to Atlas features, configuration options and operational controls. For a first Mongoose connection or a small side project, the free tier is enough. For production traffic, plan the upgrade before you need it. The tier naming matters when you read tutorials. Many older posts call the free cluster M0, and current Atlas docs still use that label while also grouping free and Flex clusters under newer service names. If a guide's console screenshots look different from yours, check the date on the guide before assuming you clicked wrong. Atlas runs the cluster in a cloud region you pick. Choosing a region close to your application or to yourself keeps latency low and matters more than the provider for a learning project. MongoDB supports AWS, Google Cloud and Azure as cluster hosts. ## Step by step: creating the cluster and connecting The current path is short: create a project, deploy a free cluster, create a database user, add your IP address, then copy the connection string from the connect dialog. Each step has a corresponding entry in the Atlas documentation and in the Atlas CLI. The numbered flow below is the one that matches current Atlas documentation, not the 2020 console screens in the video. 1. Sign in to MongoDB Atlas and create or select a project. Atlas sometimes groups a new account automatically. 2. Choose Create a cluster and pick the free tier. Select a cloud provider and a region. Name the cluster if you want a memorable label. 3. Wait for provisioning. The video quotes a few minutes, and that is still a reasonable expectation for a free cluster. 4. Create a database user with a username and a generated password. The password is shown once, so store it before closing the dialog. 5. Add your current IP address to the project IP access list. The current documentation offers this as part of the guided setup, and the Atlas CLI can do it in one command. 6. Open the cluster's connect dialog, choose Drivers or Connect to your application, and copy the connection string. 7. Replace the placeholder username and password in the string, then pass it to your driver or to Mongoose. The Atlas CLI offers a shortcut for the whole sequence. Running the setup command creates a cluster, loads sample data, adds your IP address and creates a MongoDB user in one pass, according to MongoDB's CLI documentation. That is useful when you want a reproducible environment rather than clicking through screens. ## The connection string and the allowlist, explained The connection string is a URI that carries the cluster host, an optional database name and placeholders for your username and password. The IP allowlist is a separate control that decides which network addresses may OpenAI connection at all. A typical string looks like a hostname followed by options, with the username and password inserted where the placeholders sit. The video instructs viewers to copy the string, then paste their own credentials into it before using it in code. That part of the workflow has not changed: the string is generated for you, but the credentials are yours to substitute. The allowlist is where the 2020 tutorial and current practice diverge most. The video recommends allowing access from anywhere so the database will accept connections, and says the specific IP can be tightened later. That works, but it means anyone who obtains the credentials can connect from any network. Allowlisting only your own IP is safer and takes one extra click in the current console. Two failure modes account for most first-connection errors. A missing allowlist entry produces a timeout, not an authentication error, and a password that still contains characters needing URL encoding produces an authentication failure. If the string fails, check the network access list before rewriting the password. ## How the 2020 video compares with the 2026 console The 2020 video and the current console describe the same four objects in a different order, so a reader following the video on today's Atlas will see unfamiliar screens but reach the same result. The setup flow also appears in the Atlas CLI and the Administration API, not just the web console. The API accepts a cluster creation request with the free tier instance size, which matters if you want to script environment creation for a course or a workshop. One caveat on compatibility: the video's advice to allow all IPs is a security tradeoff, not a requirement. Current documentation leads with adding your own address, and the wider setting should be temporary at most. If you are building a teaching repository, document the allowlist step explicitly rather than copying the permissive default. ## Where TypeScript and Crazystack fit in MongoDB Atlas is the hosted database layer, and TypeScript is one of several languages that connect to it. Builders who want a typed Node.js setup often pair Atlas with a starter that already wires authentication, database access and deployment, and [Crazystack TypeScript](https://crazystack.com.br) is a Brazilian starter aimed at that combination. The pairing is pragmatic: Atlas gives you the database, and a starter gives you the surrounding application shell. Neither one removes the need to understand the connection string or the allowlist, because those are properties of the database service, not of the template. If you follow the 2020 video and then move to a TypeScript stack, keep the environment variable containing the connection string out of source control. The video's advice to paste a username and password directly into code is the habit most worth leaving behind. ## FAQ - **Do I need a credit card for the MongoDB Atlas free tier?** No. The free M0 cluster can be created without billing details, and MongoDB's current documentation describes it as a development environment that does not expire. You only add payment information if you move to a paid tier. - **How long does it take to provision an Atlas cluster?** A few minutes for the free tier, which matches the wait the 2020 video quotes. The Atlas CLI setup command automates cluster creation, sample data loading, IP allowlisting and user creation in one run. - **Why does my connection time out even with the right password?** The IP allowlist is the usual cause. If your current address is not on the project's access list, Atlas refuses the network connection before authentication is attempted, so a timeout points at the network layer rather than the credentials. - **Can I have more than one free Atlas cluster?** Each Atlas project allows exactly one free cluster, per MongoDB's documentation. To run a second free database, create another project instead of another cluster in the same one. - **Is MongoDB Atlas fully open source?** No. Atlas is a managed commercial service operated by MongoDB. The MongoDB database server itself is distributed under a source-available license, and the Atlas control plane and operational features are proprietary to the service. - **What changed in MongoDB Atlas since 2020?** The console moved to a guided first-cluster flow, the CLI and Administration API became documented ways to create clusters, and free and Flex clusters were grouped under newer service naming. The core objects (cluster, user, allowlist, connection string) are unchanged. - **Should I allow access from anywhere, like the video suggests?** Only temporarily, and ideally not at all. Allowing all addresses makes the database reachable from any network once credentials leak. Adding your own address is the safer default and is what current documentation leads with. - **Which cloud region should I choose for a free Atlas cluster?** The region closest to where your application runs. For a local learning project, the nearest region to you keeps round trips small, and the provider choice among AWS, Google Cloud and Azure has little effect at this scale. - **Can the Atlas CLI replace the web console entirely?** For cluster creation, mostly yes. The setup command covers cluster creation, sample data, IP access and user creation, though you may still use the console to inspect collections and connection details, as the video shows. ## Turn your own walkthrough into a written guide The 2020 video's value is not the console screens, which have moved on. It is the sequence of decisions, from choosing a free tier to pasting a connection string, that a beginner has to make in order. Written guides preserve that sequence better than video because the reader can jump to the allowlist step without scrubbing a timeline. If you already explain technical workflows on YouTube, the same knowledge can live as an article. [Skala Blog](https://skalablog.com) transcribes a YouTube video and turns it into a structured draft you can edit before publishing, so a walkthrough recorded once can serve readers searching for the written version. [Source video](https://www.youtube.com/watch?v=upaF_Q1SXJQ)