diff --git a/boaviztapi/data/utils/complete_AWS_EC2/README.md b/boaviztapi/data/utils/complete_AWS_EC2/README.md
new file mode 100644
index 00000000..7025b7c9
--- /dev/null
+++ b/boaviztapi/data/utils/complete_AWS_EC2/README.md
@@ -0,0 +1,26 @@
+# Completing and updating Azure VMs data
+
+## Source data
+
+ 1. Go to https://instances.vantage.sh.
+ 2. Add all columns
+ 3. Click export in upper right corner, left to the search bar.
+ 4. Override vantage-export.csv
+
+## Workflow
+
+The `addData.go` script reads `vantage-export.csv`, `aws.csv` and `cpu_specs.csv`. If an instance vantage-export.csv is not found in aws.csv, this instance is added.
+
+For each instance family, if a **.metal** instance size exists, it is supposed that the main characteristics (CPU, RAM) of this instance matches the configuration of the hardware "platform" (here a platform means our best bet on the underlying hardware).
+
+If not, the largest instance size of the family is used as the underlying platform.
+
+In both cases, the platform identifier is added as a value of the platform column in `aws.csv`.
+
+Hardware configuration of each platform is described in `archetypes/server.csv`.
+
+Regarding storage, only the instances that have an explicit default storage volume in AWS documentation, with an explicit storage size, will get a storage value in aws.csv.
+
+The storage of each platform, in server.csv, is the storage of the largest instance for the related family.
+
+At this point, BoaviztAPI computes the impacts based on server.csv and aws.csv, using the [following methodology](https://doc.api.boavizta.org/Explanations/services/cloud/).