1 files changed, 48 insertions, 0 deletions
diff --git a/gems/csv_builder/README.md b/gems/csv_builder/README.md
new file mode 100644
index 00000000000..37dde4b334c
--- /dev/null
+++ b/gems/csv_builder/README.md
@@ -0,0 +1,48 @@
+# CsvBuilder
+
+## Usage
+
+Generate a CSV given a collection and a mapping.
+
+```ruby
+columns = {
+  'Title' => 'title',
+  'Comment' => 'comment',
+  'Author' => -> (post) { post.author.full_name }
+  'Created At (UTC)' => -> (post) { post.created_at&.strftime('%Y-%m-%d %H:%M:%S') }
+}
+
+CsvBuilder.new(@posts, columns).render
+```
+
+When the value of the mapping is a string, a method is called with the given name
+on the record (for example: `post.title`).
+When the value of the mapping is a lambda, it is lazily executed.
+
+It's possible to also pass ActiveRecord associations to preload when batching
+through the collection:
+
+```ruby
+CsvBuilder.new(@posts, columns, [:author, :comments]).render
+```
+
+### SingleBatch builder
+
+When the collection is an array or enumerable you can use:
+
+```ruby
+CsvBuilder::SingleBatch.new(@posts, columns).render
+```
+
+### Stream builder
+
+A stream builder uses a lazy and more efficient iterator and by default returns
+up to 100,000 records from the collection.
+
+```ruby
+CsvBuilder::Stream.new(@posts, columns).render(1_000)
+```
+
+## Development
+
+Follow the GitLab [gems development guidelines](../../doc/development/gems.md).