diff --git a/poetry.lock b/poetry.lock
index 2883895..42d9ba2 100644
--- a/poetry.lock
+++ b/poetry.lock
@@ -1,4 +1,4 @@
-# This file is automatically @generated by Poetry 2.1.1 and should not be changed by hand.
+# This file is automatically @generated by Poetry 2.1.2 and should not be changed by hand.
 
 [[package]]
 name = "ag-ui-protocol"
@@ -1293,7 +1293,6 @@ description = "Consume Server-Sent Event (SSE) messages with HTTPX."
 optional = false
 python-versions = ">=3.8"
 groups = ["main"]
-markers = "platform_system != \"Emscripten\""
 files = [
     {file = "httpx-sse-0.4.0.tar.gz", hash = "sha256:1e81a3a3070ce322add1d3529ed42eb5f70817f45ed6ec915ab753f961139721"},
     {file = "httpx_sse-0.4.0-py3-none-any.whl", hash = "sha256:f329af6eae57eaa2bdfd962b42524764af68075ea87370a2de920af5341e318f"},
@@ -1494,7 +1493,7 @@ version = "3.1.6"
 description = "A very fast and expressive template engine."
 optional = false
 python-versions = ">=3.7"
-groups = ["dev"]
+groups = ["main", "dev"]
 files = [
     {file = "jinja2-3.1.6-py3-none-any.whl", hash = "sha256:85ece4451f492d0c13c5dd7c13a64681a86afae63a5f347908daf103ce6d2f67"},
     {file = "jinja2-3.1.6.tar.gz", hash = "sha256:0137fb05990d35f1275a587e9aee6d56da821fc83491a0fb838183be43f66d6d"},
@@ -1700,7 +1699,7 @@ version = "3.0.2"
 description = "Safely add untrusted strings to HTML/XML markup."
 optional = false
 python-versions = ">=3.9"
-groups = ["dev"]
+groups = ["main", "dev"]
 files = [
     {file = "MarkupSafe-3.0.2-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:7e94c425039cde14257288fd61dcfb01963e658efbc0ff54f5306b06054700f8"},
     {file = "MarkupSafe-3.0.2-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:9e2d922824181480953426608b81967de705c3cef4d1af983af849d7bd619158"},
@@ -3995,4 +3994,4 @@ type = ["pytest-mypy"]
 [metadata]
 lock-version = "2.1"
 python-versions = ">=3.12,<4"
-content-hash = "3b07173aee03bc879861259f0087d89ceaed03d405344a6d8cc0381b7772d398"
+content-hash = "153e9037ab8c82802d4b638f6b043ae346581db66a478cf481d2fc8a60a3427c"
diff --git a/pyproject.toml b/pyproject.toml
index 5f3ff96..cbd0f7a 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -37,7 +37,8 @@ dependencies = [
     "python-gitlab>=5.1.0,<6.0.0",
     "rich>=13.9.4,<14.0.0",
     "pygithub>=2.6.1,<3.0.0",
-    "httpx (>=0.28.1,<0.29.0)"
+    "httpx (>=0.28.1,<0.29.0)",
+    "jinja2 (>=3.1.6,<4.0.0)"
 ]
 
 [tool.poetry.group.dev.dependencies]
diff --git a/src/lgtm_ai/formatters/markdown.py b/src/lgtm_ai/formatters/markdown.py
index f97ae3d..dced7b5 100644
--- a/src/lgtm_ai/formatters/markdown.py
+++ b/src/lgtm_ai/formatters/markdown.py
@@ -1,143 +1,92 @@
-import textwrap
+import pathlib
+from typing import ClassVar
 
-from lgtm_ai.ai.schemas import PublishMetadata, Review, ReviewComment, ReviewGuide, ReviewScore
+from jinja2 import Environment, FileSystemLoader
+from lgtm_ai.ai.schemas import PublishMetadata, Review, ReviewComment, ReviewGuide
 from lgtm_ai.formatters.base import Formatter
 from lgtm_ai.formatters.constants import CATEGORY_MAP, SCORE_MAP, SEVERITY_MAP
 from pydantic_ai.usage import Usage
 
 
 class MarkDownFormatter(Formatter[str]):
-    def format_review_summary_section(self, review: Review, comments: list[ReviewComment] | None = None) -> str:
-        header = textwrap.dedent(f"""
-        ## 🦉 lgtm Review
-
-        > **Score:** {self._format_score(review.review_response.score)}
-
-        ### 🔍 Summary
-
-        """)
-        summary = header + review.review_response.summary
-        if comments:
-            summary += f"\n\n{self.format_review_comments_section(comments)}"
+    REVIEW_SUMMARY_TEMPLATE: ClassVar[str] = "review_summary.md.j2"
+    REVIEW_COMMENTS_SECTION_TEMPLATE: ClassVar[str] = "review_comments_section.md.j2"
+    REVIEW_COMMENT_TEMPLATE: ClassVar[str] = "review_comment.md.j2"
+    REVIEW_GUIDE_TEMPLATE: ClassVar[str] = "review_guide.md.j2"
+    SNIPPET_TEMPLATE: ClassVar[str] = "snippet.md.j2"
+    USAGES_SUMMARY_TEMPLATE: ClassVar[str] = "usages_summary.md.j2"
+    METADATA_TEMPLATE: ClassVar[str] = "metadata.md.j2"
+
+    def __init__(self) -> None:
+        template_dir = pathlib.Path(__file__).parent / "templates"
+        self.env = Environment(loader=FileSystemLoader(template_dir), autoescape=True)
 
-        summary += self._format_metadata(review.metadata)
-        return summary
+    def format_review_summary_section(self, review: Review, comments: list[ReviewComment] | None = None) -> str:
+        template = self.env.get_template(self.REVIEW_SUMMARY_TEMPLATE)
+        comments_section = self.format_review_comments_section(comments or [])
+        metadata = self._format_metadata(review.metadata)
+        return template.render(
+            score=review.review_response.score,
+            score_icon=SCORE_MAP[review.review_response.score],
+            summary=review.review_response.summary,
+            comments_section=comments_section,
+            metadata=metadata,
+        )
 
     def format_review_comments_section(self, comments: list[ReviewComment]) -> str:
         if not comments:
             return ""
-        lines = ["**Specific Comments:**"]
-        for comment in comments:
-            lines.append(f"- {self.format_review_comment(comment, with_footer=False)}")
-        return "\n\n".join(lines)
+        template = self.env.get_template(self.REVIEW_COMMENTS_SECTION_TEMPLATE)
+        rendered_comments = [self.format_review_comment(comment, with_footer=False) for comment in comments]
+        return template.render(comments=rendered_comments)
 
     def format_review_comment(self, comment: ReviewComment, *, with_footer: bool = True) -> str:
-        header_section = "\n\n".join(
-            [
-                f"#### 🦉 {CATEGORY_MAP[comment.category]} {comment.category}",
-                f"> **Severity:** {comment.severity} {SEVERITY_MAP[comment.severity]}",
-            ]
+        template = self.env.get_template(self.REVIEW_COMMENT_TEMPLATE)
+        header_category = CATEGORY_MAP[comment.category]
+        severity_icon = SEVERITY_MAP[comment.severity]
+        snippet = self._format_snippet(comment) if comment.quote_snippet else None
+        return template.render(
+            category=header_category,
+            category_key=comment.category,
+            severity=comment.severity,
+            severity_icon=severity_icon,
+            snippet=snippet,
+            comment=comment.comment,
+            with_footer=with_footer,
+            new_path=comment.new_path,
+            line_number=comment.line_number,
+            relative_line_number=comment.relative_line_number,
         )
-        comment_section = (
-            f"\n{self._format_snippet(comment)}\n{comment.comment}" if comment.quote_snippet else comment.comment
-        )
-
-        footer_section = (
-            textwrap.dedent(f"""
-
-        <details><summary>More information about this comment</summary>
-
-        - **File**: `{comment.new_path}`
-        - **Line**: `{comment.line_number}`
-        - **Relative line**: `{comment.relative_line_number}`
-
-        </details>
-        """)
-            if with_footer
-            else ""
-        )
-
-        return f"{header_section}\n\n{comment_section}\n\n{footer_section}"
 
     def format_guide(self, guide: ReviewGuide) -> str:
-        header = textwrap.dedent("""
-        ## 🦉 lgtm Reviewer Guide
-
-        """)
-
-        summary = guide.guide_response.summary
-        # Format key changes as a markdown table
-        key_changes = ["| File Name | Description |", "| ---- | ---- |"] + [
-            f"| {change.file_name} | {change.description} |" for change in guide.guide_response.key_changes
-        ]
-
-        # Format checklist items as a checklist
-        checklist = [f"- [ ] {item.description}" for item in guide.guide_response.checklist]
-
-        # Format references as a list
-        if guide.guide_response.references:
-            references = [f"- [{item.title}]({item.url})" for item in guide.guide_response.references]
-        else:
-            references = []
-
-        # Combine all sections
-
-        summary = (
-            header
-            + "### 🔍 Summary\n\n"
-            + summary
-            + "\n\n### 🔑 Key Changes\n\n"
-            + "\n".join(key_changes)
-            + "\n\n### ✅ Reviewer Checklist\n\n"
-            + "\n".join(checklist)
+        template = self.env.get_template(self.REVIEW_GUIDE_TEMPLATE)
+        key_changes = guide.guide_response.key_changes
+        checklist = guide.guide_response.checklist
+        references = guide.guide_response.references
+        metadata = self._format_metadata(guide.metadata)
+        return template.render(
+            summary=guide.guide_response.summary,
+            key_changes=key_changes,
+            checklist=checklist,
+            references=references,
+            metadata=metadata,
         )
-        if references:
-            summary += "\n\n### 📚 References\n\n" + "\n".join(references)
-
-        summary += self._format_metadata(guide.metadata)
-        return summary
-
-    def _format_score(self, score: ReviewScore) -> str:
-        return f"{score} {SCORE_MAP[score]}"
 
     def _format_snippet(self, comment: ReviewComment) -> str:
-        return f"\n\n```{comment.programming_language.lower()}\n{comment.quote_snippet}\n```\n\n"
+        template = self.env.get_template(self.SNIPPET_TEMPLATE)
+        return template.render(language=comment.programming_language.lower(), snippet=comment.quote_snippet)
 
     def _format_usages_summary(self, usages: list[Usage]) -> str:
-        formatted_usage_calls = []
-        for i, usage in enumerate(usages):
-            formatted_usage_calls += [self._format_usage_call_collapsible(usage, i)]
-
-        return f"""
-        <details><summary>Usage summary</summary>
-        {"\n".join(formatted_usage_calls)}
-        **Total tokens**: `{sum([usage.total_tokens or 0 for usage in usages])}`
-        </details>
-        """
-
-    def _format_usage_call_collapsible(self, usage: Usage, index: int) -> str:
-        return f"""
-        <details><summary>Call {index + 1}</summary>
-
-        - **Request count**: `{usage.requests}`
-        - **Request tokens**: `{usage.request_tokens}`
-        - **Response tokens**: `{usage.response_tokens}`
-        - **Total tokens**: `{usage.total_tokens}`
-        </details>
-        """
+        template = self.env.get_template(self.USAGES_SUMMARY_TEMPLATE)
+        total_tokens = sum(usage.total_tokens or 0 for usage in usages)
+        return template.render(usages=usages, total_tokens=total_tokens)
 
     def _format_metadata(self, metadata: PublishMetadata) -> str:
-        return textwrap.dedent(f"""
-
-        <details><summary>More information</summary>
-
-        - **Id**: `{metadata.uuid}`
-        - **Model**: `{metadata.model_name}`
-        - **Created at**: `{metadata.created_at}`
-
-        {self._format_usages_summary(metadata.usages)}
-
-        > See the [📚 lgtm-ai repository](https://github.com/elementsinteractive/lgtm-ai) for more information about lgtm.
-
-        </details>
-        """)
+        template = self.env.get_template(self.METADATA_TEMPLATE)
+        usages_summary = self._format_usages_summary(metadata.usages)
+        return template.render(
+            uuid=metadata.uuid,
+            model_name=metadata.model_name,
+            created_at=metadata.created_at,
+            usages_summary=usages_summary,
+        )
diff --git a/src/lgtm_ai/formatters/templates/metadata.md.j2 b/src/lgtm_ai/formatters/templates/metadata.md.j2
new file mode 100644
index 0000000..a4749cb
--- /dev/null
+++ b/src/lgtm_ai/formatters/templates/metadata.md.j2
@@ -0,0 +1,14 @@
+<details><summary>More information</summary>
+
+- **Id**: `{{ uuid }}`
+- **Model**: `{{ model_name }}`
+- **Created at**: `{{ created_at }}`
+
+
+{{ usages_summary | safe }}
+
+
+> See the [📚 lgtm-ai repository](https://github.com/elementsinteractive/lgtm-ai) for more information about lgtm.
+
+</details>
+
diff --git a/src/lgtm_ai/formatters/templates/review_comment.md.j2 b/src/lgtm_ai/formatters/templates/review_comment.md.j2
new file mode 100644
index 0000000..c4012a2
--- /dev/null
+++ b/src/lgtm_ai/formatters/templates/review_comment.md.j2
@@ -0,0 +1,18 @@
+#### 🦉 {{ category }} {{ category_key }}
+
+> **Severity:** {{ severity }} {{ severity_icon }}
+
+{% if snippet %}
+{{ snippet | safe }}
+
+{% endif %}
+{{ comment | safe }}
+
+{% if with_footer %}
+<details><summary>More information about this comment</summary>
+
+- **File**: `{{ new_path }}`
+- **Line**: `{{ line_number }}`
+- **Relative line**: `{{ relative_line_number }}`
+</details>
+{% endif %}
\ No newline at end of file
diff --git a/src/lgtm_ai/formatters/templates/review_comments_section.md.j2 b/src/lgtm_ai/formatters/templates/review_comments_section.md.j2
new file mode 100644
index 0000000..8b85a8b
--- /dev/null
+++ b/src/lgtm_ai/formatters/templates/review_comments_section.md.j2
@@ -0,0 +1,5 @@
+{% if comments %}**Specific Comments:**
+
+{% for comment in comments %}- {{ comment | safe }}
+{% endfor %}
+{% endif %}
\ No newline at end of file
diff --git a/src/lgtm_ai/formatters/templates/review_guide.md.j2 b/src/lgtm_ai/formatters/templates/review_guide.md.j2
new file mode 100644
index 0000000..424d977
--- /dev/null
+++ b/src/lgtm_ai/formatters/templates/review_guide.md.j2
@@ -0,0 +1,31 @@
+
+## 🦉 lgtm Reviewer Guide
+
+### 🔍 Summary
+
+{{ summary | safe }}
+
+### 🔑 Key Changes
+
+| File Name | Description |
+| ---- | ---- |
+{% for change in key_changes -%}
+| {{ change.file_name }} | {{ change.description }} |
+{% endfor %}
+
+### ✅ Reviewer Checklist
+
+{% for item in checklist %}
+- [ ] {{ item.description }}
+{% endfor %}
+
+{% if references %}
+### 📚 References
+{% for ref in references %}
+- [{{ ref.title }}]({{ ref.url }})
+{% endfor %}
+{% endif %}
+
+{{ metadata | safe }}
+
+
diff --git a/src/lgtm_ai/formatters/templates/review_summary.md.j2 b/src/lgtm_ai/formatters/templates/review_summary.md.j2
new file mode 100644
index 0000000..d241c86
--- /dev/null
+++ b/src/lgtm_ai/formatters/templates/review_summary.md.j2
@@ -0,0 +1,14 @@
+
+## 🦉 lgtm Review
+
+> **Score:** {{ score }} {{ score_icon }}
+
+### 🔍 Summary
+
+{{ summary | safe }}
+
+{% if comments_section %}
+{{ comments_section | safe}}
+{% endif %}
+
+{{ metadata | safe }}
diff --git a/src/lgtm_ai/formatters/templates/snippet.md.j2 b/src/lgtm_ai/formatters/templates/snippet.md.j2
new file mode 100644
index 0000000..f058406
--- /dev/null
+++ b/src/lgtm_ai/formatters/templates/snippet.md.j2
@@ -0,0 +1,3 @@
+```{{ language }}
+{{ snippet | safe }}
+```
diff --git a/src/lgtm_ai/formatters/templates/usages_summary.md.j2 b/src/lgtm_ai/formatters/templates/usages_summary.md.j2
new file mode 100644
index 0000000..a3d3ee7
--- /dev/null
+++ b/src/lgtm_ai/formatters/templates/usages_summary.md.j2
@@ -0,0 +1,14 @@
+<details><summary>Usage summary</summary>
+{% for usage in usages %}
+
+<details><summary>Call {{ loop.index }}</summary>
+
+- **Request count**: `{{ usage.requests }}`
+- **Request tokens**: `{{ usage.request_tokens }}`
+- **Response tokens**: `{{ usage.response_tokens }}`
+- **Total tokens**: `{{ usage.total_tokens }}`
+</details>
+
+{% endfor %}
+**Total tokens**: `{{ total_tokens }}`
+</details>
diff --git a/tests/formatters/test_markdown.py b/tests/formatters/test_markdown.py
index d30db78..90ef930 100644
--- a/tests/formatters/test_markdown.py
+++ b/tests/formatters/test_markdown.py
@@ -44,6 +44,8 @@ def test_format_summary_section(self) -> None:
             "",
             "summary",
             "",
+            "",
+            "",
             "<details><summary>More information</summary>",
             "",
             "- **Id**: `fb64cb958fcf49219545912156e0a4a0`",
@@ -53,6 +55,7 @@ def test_format_summary_section(self) -> None:
             "",
             "<details><summary>Usage summary</summary>",
             "",
+            "",
             "<details><summary>Call 1</summary>",
             "",
             "- **Request count**: `1`",
@@ -62,6 +65,7 @@ def test_format_summary_section(self) -> None:
             "</details>",
             "",
             "",
+            "",
             "<details><summary>Call 2</summary>",
             "",
             "- **Request count**: `1`",
@@ -71,6 +75,7 @@ def test_format_summary_section(self) -> None:
             "</details>",
             "",
             "",
+            "",
             "<details><summary>Call 3</summary>",
             "",
             "- **Request count**: `1`",
@@ -79,6 +84,7 @@ def test_format_summary_section(self) -> None:
             "- **Total tokens**: `300`",
             "</details>",
             "",
+            "",
             "**Total tokens**: `900`",
             "</details>",
             "",
@@ -148,20 +154,35 @@ def test_format_comments_section_several_comments(self) -> None:
 
         expected = [
             "**Specific Comments:**",
+            "",
             "- #### 🦉 🧪 Testing",
+            "",
             "> **Severity:** HIGH 🔴",
+            "",
+            "",
             "comment 2",
             "",
+            "",
             "- #### 🦉 🧪 Testing",
+            "",
             "> **Severity:** MEDIUM 🟡",
+            "",
+            "",
             "comment 3",
             "",
+            "",
             "- #### 🦉 🎯 Correctness",
+            "",
             "> **Severity:** LOW 🔵",
+            "",
+            "",
             "comment 1",
             "",
+            "",
+            "",
+            "",
         ]
-        assert self.formatter.format_review_comments_section(review.review_response.comments) == "\n\n".join(expected)
+        assert self.formatter.format_review_comments_section(review.review_response.comments).split("\n") == expected
 
     def test_format_comment_with_snippet(self) -> None:
         review = Review(
@@ -189,16 +210,27 @@ def test_format_comment_with_snippet(self) -> None:
 
         expected = [
             "#### 🦉 🎯 Correctness",
+            "",
             "> **Severity:** LOW 🔵",
             "",
-            "\n```python\nprint('Hello World')\n```",
-            "\ncomment",
+            "",
+            "```python",
+            "print('Hello World')",
+            "```",
+            "",
+            "",
+            "comment",
+            "",
             "",
             "<details><summary>More information about this comment</summary>",
-            "- **File**: `new_path`\n- **Line**: `1`\n- **Relative line**: `1`",
-            "</details>\n",
+            "",
+            "- **File**: `new_path`",
+            "- **Line**: `1`",
+            "- **Relative line**: `1`",
+            "</details>",
+            "",
         ]
-        assert self.formatter.format_review_comment(review.review_response.comments[0]) == "\n\n".join(expected)
+        assert self.formatter.format_review_comment(review.review_response.comments[0]).split("\n") == expected
 
     def test_format_guide(self) -> None:
         guide = ReviewGuide(
@@ -228,27 +260,62 @@ def test_format_guide(self) -> None:
                 spec=PublishMetadata,
             ),
         )
-        assert self.formatter.format_guide(guide).split("\n\n") == [
-            "\n## 🦉 lgtm Reviewer Guide",
+        assert self.formatter.format_guide(guide).split("\n") == [
+            "",
+            "## 🦉 lgtm Reviewer Guide",
+            "",
             "### 🔍 Summary",
+            "",
             "summary",
+            "",
             "### 🔑 Key Changes",
-            "| File Name | Description |\n| ---- | ---- |\n| foo.py | description |\n| bar.py | description |",
+            "",
+            "| File Name | Description |",
+            "| ---- | ---- |",
+            "| foo.py | description |",
+            "| bar.py | description |",
+            "",
+            "",
             "### ✅ Reviewer Checklist",
+            "",
+            "",
             "- [ ] item 1",
+            "",
+            "",
+            "",
             "### 📚 References",
+            "",
             "- [title](https://example.com)",
+            "",
+            "",
+            "",
             "<details><summary>More information</summary>",
-            "- **Id**: `fb64cb958fcf49219545912156e0a4a0`\n- **Model**: `whatever`\n- **Created at**: `2025-05-15T09:43:01.654374+00:00`",
-            "\n<details><summary>Usage summary</summary>",
+            "",
+            "- **Id**: `fb64cb958fcf49219545912156e0a4a0`",
+            "- **Model**: `whatever`",
+            "- **Created at**: `2025-05-15T09:43:01.654374+00:00`",
+            "",
+            "",
+            "<details><summary>Usage summary</summary>",
+            "",
+            "",
             "<details><summary>Call 1</summary>",
-            "- **Request count**: `1`\n"
-            "- **Request tokens**: `200`\n"
-            "- **Response tokens**: `100`\n"
-            "- **Total tokens**: `300`\n"
+            "",
+            "- **Request count**: `1`",
+            "- **Request tokens**: `200`",
+            "- **Response tokens**: `100`",
+            "- **Total tokens**: `300`",
             "</details>",
-            "**Total tokens**: `300`\n</details>",
-            "\n"
+            "",
+            "",
+            "**Total tokens**: `300`",
+            "</details>",
+            "",
+            "",
             "> See the [📚 lgtm-ai repository](https://github.com/elementsinteractive/lgtm-ai) for more information about lgtm.",
-            "</details>\n",
+            "",
+            "</details>",
+            "",
+            "",
+            "",
         ]