diff --git a/.gitignore b/.gitignore
index 6a014f5..7e00121 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,5 +1,6 @@
 .sass-cache
 .DS_Store
+.my.cnf
 
 # github premade rules
 *.py[cod]
diff --git a/bitshift/database.py b/bitshift/database.py
index 36b984e..647fe55 100644
--- a/bitshift/database.py
+++ b/bitshift/database.py
@@ -3,19 +3,18 @@ Module with classes and functions to handle communication with the MySQL
 database backend, which manages the search index.
 """
 
+import mmh3
 import oursql
 
 class Database(object):
     """Represents the MySQL database."""
 
     def __init__(self):
-        pass
+        self._connect()
 
     def _connect(self):
-        pass
-
-    def _create(self):
-        pass
+        """Establish a connection to the database."""
+        self._conn = oursql.connect()
 
     def search(self, query):
         """
@@ -36,4 +35,5 @@ class Database(object):
         :param codelet: The codelet to insert.
         :type codelet: :py:class:`.Codelet`
         """
+        # code_hash = mmh3.hash64(codelet.code)[0]
         pass
diff --git a/schema.sql b/schema.sql
new file mode 100644
index 0000000..3cb915c
--- /dev/null
+++ b/schema.sql
@@ -0,0 +1,23 @@
+CREATE DATABASE bitshift DEFAULT CHARACTER SET utf8 COLLATE utf8_unicode_ci;
+USE `bitshift`;
+
+CREATE TABLE codelets (
+    `codelet_id` BIGINT UNSIGNED NOT NULL AUTO_INCREMENT UNIQUE,
+    `codelet_name` VARCHAR(512) NOT NULL,
+    `codelet_code_id` BIGINT UNSIGNED NOT NULL,
+    `codelet_lang` SMALLINT UNSIGNED DEFAULT NULL,
+    `codelet_origin` TINYINT UNSIGNED DEFAULT NULL,
+    `codelet_url` VARCHAR(512) NOT NULL,
+    `codelet_date_created` DATETIME DEFAULT NULL,
+    `codelet_date_modified` DATETIME DEFAULT NULL,
+    PRIMARY KEY (`codelet_id`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci;
+
+CREATE TABLE code (
+    `code_id` BIGINT UNSIGNED NOT NULL AUTO_INCREMENT UNIQUE,
+    `code_hash` BIGINT NOT NULL,
+    `code_code` MEDIUMTEXT NOT NULL,
+    PRIMARY KEY (`code_id`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci;
+
+-- separate tables: authors, symbols, caches, search indices
diff --git a/setup.py b/setup.py
index 5fa1189..97441b7 100644
--- a/setup.py
+++ b/setup.py
@@ -4,8 +4,9 @@ setup(
     name = "bitshift",
     version = "0.1",
     packages = find_packages(),
-    install_requires = ["Flask>=0.10.1", "pygments>=1.6", "requests>=2.2.0",
-                        "BeautifulSoup>=3.2.1", "oursql>=0.9.3.1"],
+    install_requires = [
+        "Flask>=0.10.1", "pygments>=1.6", "requests>=2.2.0",
+        "BeautifulSoup>=3.2.1", "oursql>=0.9.3.1", "mmh3>=2.3"],
     author = "Benjamin Attal, Ben Kurtovic, Severyn Kozak",
     license = "MIT",
     url = "https://github.com/earwig/bitshift"