diff --git a/README.md b/README.md
new file mode 100644
index 0000000..ba8a4f5
--- /dev/null
+++ b/README.md
@@ -0,0 +1,43 @@
+<div align="center">
+<h1>mCaptcha Survey</h1>
+<p>
+
+**Performance statics survey runner**
+
+</p>
+
+[![Docker](https://img.shields.io/docker/pulls/mcaptcha/survey)](https://hub.docker.com/r/mcaptcha/survey)
+[![Build](https://github.com/mCaptcha/survey/actions/workflows/linux.yml/badge.svg)](https://github.com/mCaptcha/survey/actions/workflows/linux.yml)
+[![dependency status](https://deps.rs/repo/github/mCaptcha/survey/status.svg)](https://deps.rs/repo/github/mCaptcha/survey)
+[![codecov](https://codecov.io/gh/mCaptcha/survey/branch/master/graph/badge.svg)](https://codecov.io/gh/mCaptcha/survey)
+
+</div>
+
+## Why
+
+[mCaptcha](https://mcaptcha.org) is a
+[proof-of-work](https://en.wikipedia.org/wiki/Proof_of_work) based
+CAPTCHA system. Its effectiveness depends on an accurate and
+time-relevant proof-of-work difficulty setting. If it is too high, it
+could end up
+[DoS-ing](https://en.wikipedia.org/wiki/Denial-of-service_attack) the
+underlying service that it is supposed to protect and if it is too low,
+the protection offered will be ineffective.
+
+In order to select the right difficulty level, mCaptcha admins would
+require knowledge about current performance benchmarks on a large
+variety of devices that are currently on the internet.
+
+
+## What
+
+This program runs a mCaptcha benchmarks on user devices and collects
+fully anonymous(only device statics are stored) performance statics,
+that are transparently made available to everyone free of charge.
+mCaptcha admins are kindly requested to refer to the benchmarks
+published to fine-tune their CAPTCHA deployment.
+
+
+## What data do you collect?
+
+TODO: run program, record and share actual network traffic logs