diff --git a/src/docgen/generators/markdown.py b/src/docgen/generators/markdown.py
new file mode 100644
index 0000000..6cb2adb
--- /dev/null
+++ b/src/docgen/generators/markdown.py
@@ -0,0 +1,50 @@
+"""Markdown documentation generator."""
+
+from pathlib import Path
+from jinja2 import Environment, FileSystemLoader
+from docgen.models import DocConfig, Endpoint
+from docgen.generators import BaseGenerator
+
+
+class MarkdownGenerator(BaseGenerator):
+    """Generator for Markdown documentation."""
+
+    def __init__(self, config: DocConfig = None):
+        super().__init__(config)
+        template_dir = Path(__file__).parent.parent / "templates"
+        self.env = Environment(loader=FileSystemLoader(template_dir))
+
+    def generate(self, endpoints: list[Endpoint], output_dir: Path) -> Path:
+        """Generate Markdown documentation."""
+        output_dir = self._ensure_output_dir(output_dir)
+
+        template = self.env.get_template("markdown.md.jinja2")
+        grouped = self._group_endpoints(endpoints)
+
+        rendered = template.render(
+            title=self.config.title,
+            description=self.config.description,
+            version=self.config.version,
+            endpoints=endpoints,
+            grouped_endpoints=grouped,
+        )
+
+        readme_path = output_dir / "README.md"
+        readme_path.write_text(rendered)
+
+        return readme_path
+
+    def _group_endpoints(self, endpoints: list[Endpoint]) -> dict[str, list[Endpoint]]:
+        """Group endpoints by tags or path prefixes."""
+        grouped = {}
+        for endpoint in endpoints:
+            if endpoint.tags:
+                tag = endpoint.tags[0] if endpoint.tags else "Other"
+            else:
+                parts = endpoint.path.strip("/").split("/")
+                tag = parts[0] if parts else "Other"
+            
+            if tag not in grouped:
+                grouped[tag] = []
+            grouped[tag].append(endpoint)
+        return dict(sorted(grouped.items()))