diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..d8e2e9e
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,4 @@
+*.pyc
+.vs/
+tmp/
+*.tsv
diff --git a/.gitmodules b/.gitmodules
new file mode 100644
index 0000000..ba76e16
--- /dev/null
+++ b/.gitmodules
@@ -0,0 +1,9 @@
+[submodule "raw/eng2tamildictionary"]
+	path = raw/eng2tamildictionary
+	url = https://github.com/linuxkathirvel/eng2tamildictionary
+[submodule "raw/english-tamil-dictionary-api"]
+	path = raw/english-tamil-dictionary-api
+	url = https://github.com/abuvanth/english-tamil-dictionary-api
+[submodule "raw/dictionary"]
+	path = raw/dictionary
+	url = https://github.com/sathia27/dictionary
diff --git a/consolidate.py b/consolidate.py
new file mode 100644
index 0000000..46ddc03
--- /dev/null
+++ b/consolidate.py
@@ -0,0 +1,62 @@
+from tqdm import tqdm
+import re
+import pandas as pd
+
+en_lang_pairs_count = {}
+
+def add_pair(en_word, lang_word):
+    if type(en_word) is not str or type(lang_word) is not str:
+        return
+    en_word = en_word.replace('"', '').strip().lower().rstrip('.')
+    lang_word = lang_word.replace('"', '').strip().rstrip('.')
+    key = (en_word, lang_word)
+    if len(key[0])<2 or len(key[1])<2:
+        return
+    if key in en_lang_pairs_count:
+        en_lang_pairs_count[key] += 1
+    else:
+        en_lang_pairs_count[key] = 1
+    return
+
+###################################################################
+
+# https://github.com/linuxkathirvel/eng2tamildictionary
+
+import json
+data = json.load(open('raw/eng2tamildictionary/dictionary.json', encoding="utf-8"))
+
+for record in tqdm(data, desc='eng2tamildictionary'):
+    try:
+        en_word, lang_words = record["eng"], record["tamil"]
+    except:
+        if 'word_list' in record:
+            continue
+        raise
+    
+    en_word = re.sub('\([^)]*?\)', '', en_word) # Remove brackets
+    lang_words = re.sub('\([^)]*?\)', '', lang_words) # Remove brackets
+    lang_words = re.sub("[A-Za-z]\.?", ' ', lang_words) # Remove English words from tamil
+    lang_words = re.sub("-[1-9]", ' ', lang_words) # Remove numbered bullets
+    for lang_word in lang_words.split(','):
+        add_pair(en_word, lang_word)
+
+# https://github.com/abuvanth/english-tamil-dictionary-api
+# Same as above, so ignore.
+
+# https://github.com/sathia27/dictionary
+# Bad format. TODO: Clean and parse
+
+# import sqlite3, os
+# con = sqlite3.connect('raw/dictionary/word.db', isolation_level=None, detect_types=sqlite3.PARSE_COLNAMES)
+
+# df = pd.read_sql_query("SELECT * FROM words", con)
+# os.makedirs('tmp', exist_ok=True)
+# df.to_csv("tmp/dict.csv")
+
+## WRITE
+
+out = open('consolidated.tsv', 'w', encoding='utf-8')
+out.write(f"ENG\tLANG\tCOUNT\n")
+for (en_word, lang_word), count in en_lang_pairs_count.items():
+    out.write(f"{en_word}\t{lang_word}\t{count}\n")
+out.close()
diff --git a/raw/dictionary b/raw/dictionary
new file mode 160000
index 0000000..6f50493
--- /dev/null
+++ b/raw/dictionary
@@ -0,0 +1 @@
+Subproject commit 6f504935d6fb3b1c0727018159fef9fe65a33201
diff --git a/raw/eng2tamildictionary b/raw/eng2tamildictionary
new file mode 160000
index 0000000..a4e0695
--- /dev/null
+++ b/raw/eng2tamildictionary
@@ -0,0 +1 @@
+Subproject commit a4e06953d6b4d25cb91a34fbf0fa91561e6241dd
diff --git a/raw/english-tamil-dictionary-api b/raw/english-tamil-dictionary-api
new file mode 160000
index 0000000..6c2ec06
--- /dev/null
+++ b/raw/english-tamil-dictionary-api
@@ -0,0 +1 @@
+Subproject commit 6c2ec06492b4680ba6c95d8449a91dd3a2b43251