diff --git a/README.md b/README.md
index 11e1fb04ece530362c3e84b68f4851da5a5ca3e1..a8bf41e90ca93c6058c4de5a3278c16a9e418003 100644
--- a/README.md
+++ b/README.md
@@ -1,2 +1,14 @@
 ## About
 
+The idea of this project was, given a source code, to generate new source codes following the same structure as the dataset. For that, I used a character-to-character architecture powered by an LSTM.
+
+
+## Disclaimer
+
+The network was tested on a small dataset and may overfit. A sample of the dataset is available [here](dataset/source-code.txt).
+
+
+## To do
+
+- [ ] Explain the source code in detail.
+- [ ] Add a Jupyter Notebook.
\ No newline at end of file