About & method

Built to tell you when the answer is no.

microfrontends.dev is a free, interactive guide for frontend engineers deciding whether their team needs micro frontends, and which approach fits if it does. It has no sponsors, no accounts and nothing to sell, so it can say “keep your monolith” when that’s the right call.

Who it’s for

One lesson, three depths

SimpleNew to this

The core idea of each step in plain words, with a demo where one helps. No architecture background needed.

DetailedShipped a few

Adds how the idea plays out in practice: shared builds and releases, team boundaries, and what the sources say.

Deep diveArchitecting one

Adds the trade-offs behind the decision, such as pipelines, governance, performance budgets and migration.

Principles

What we hold ourselves to

Honest by defaultEvery approach lists its costs next to its benefits. “Don’t do this” is a valid result.
Learn by doingLessons and approach pages come with demos you can change, so you can see an idea work, not just read about it.
Every claim citedEach lesson lists its sources, and every comparison rating has a reason.
Nothing storedNo accounts, no progress tracking. Shared results live in the link. The only thing you can choose to share is an email address, for project updates. Privacy
Accessible firstEvery demo works by keyboard and touch, and every animation respects your reduced-motion setting.

Practising what it preaches

Why this site isn’t built with micro frontends

This site is built and maintained by one small team, ships as plain static pages, and has no need for independent deploys or several frameworks. By its own reasoning that is a case for a single, well-structured app, so that is what it is: a live example of the “you probably don’t need it” outcome.

Take the decision check

How ratings work

A 1–5 judgement, always with a reason

Ratings compare approaches for a typical team. They are editorial judgements built from the sources we cite, not benchmarks, and every rating comes with a one-sentence reason and a source. See them side by side in the comparison matrix.

Try the scale
1 · Works against youThe approach makes this criterion harder than the simplest alternative.e.g. npm packages · tech-agnostic
2 · WeakPossible, but you will fight the approach to get there.e.g. Module Federation · tech-agnostic
3 · WorkableFine with deliberate effort and the right tooling.e.g. Native Federation · runtime performance
4 · StrongA clear strength of the approach.e.g. single-spa · team independence
5 · Best in classThis is what the approach is built for.e.g. iframes · team independence
Team independence
Can teams build and deploy without coordinating?
Runtime performance
Impact on bundle size, requests and time to interactive.
Complexity
Setup and ongoing operational cost. Lower is better.
SSR / SEO support
How well it works with server rendering.
Tech-agnostic
Can different frameworks coexist?
Isolation
Style, script and failure isolation.
UX consistency
How easy it is to keep a consistent experience.
Tooling maturity
Ecosystem, docs and community.

How the decision check works

Transparent weights, no black box

  1. 10 questionsAbout your teams, releases, frameworks, performance needs and constraints.
  2. A readiness scoreAnswers add up to a score out of 100 that says how ready your team is for micro frontends.
  3. Rules can override itA few hard rules, such as “single team, single framework”, decide the outcome whatever the score.
  4. Reasons you can checkEach approach is scored on how well it fits, and every answer shows how it affected the result.

Every weight, rule and band is published in how the decision check works. Disagree with one? Open an issue on GitHub.

Policies

Rules we write by

Editorial rules
  1. Every factual claim about a tool, its behavior or its status links to a primary source: official documentation, a specification, or the author’s own publication.
  2. No invented statistics, quotes or case studies. A claim that can’t be traced to a source is removed.
  3. Opinions and recommendations are labelled as the site’s view and kept apart from facts.
  4. Every comparison rating comes with a one-sentence reason and a source.

Every source the site cites is listed on Sources & method, with the pages that rely on it.

Neutral and free

No ads, no paid content and no sponsors, and no ties to any tool, framework or vendor. Every approach gets the same template and the same criteria, including when to avoid it.

Privacy

No accounts, no cookies and no analytics. Your decision check answers live only in the result link you choose to share, and we only have your email address if you subscribe. Read the privacy page.

For AI assistants

llms.txt lists every page with a one-line summary, for language models and the tools built on them. Every page also has a Markdown version at its URL with .md added, andllms-full.txt holds the full text of every page in one file.

Found something wrong?

Corrections are the most useful contribution. Every lesson has a “Report an issue on GitHub” link that fills in the page for you, and the site’s content and code live in the GitHub repository, where every change is proposed as a pull request and reviewed before it goes live.

Open an issue on GitHub →