diff --git a/app.py b/app.py
index ed7134f..7e16cdb 100644
--- a/app.py
+++ b/app.py
@@ -262,11 +262,18 @@ def _split_thinking(text: str) -> tuple[str, str]:
 
 def _md_to_pdf(title: str, content_md: str) -> bytes:
     """Render Markdown content to a PDF byte string."""
+
+    def _l1(text: str) -> str:
+        """Sanitize text to Latin-1; fpdf2 built-in fonts require it."""
+        return text.encode("latin-1", errors="replace").decode("latin-1")
+
+    _title = _l1(title[:100])
+
     class PDF(fpdflib.FPDF):
         def header(self):
             self.set_font("Helvetica", "B", 9)
             self.set_text_color(100, 100, 100)
-            self.cell(0, 6, title, align="L")
+            self.cell(0, 6, _title, align="L")
             self.ln(2)
             self.set_draw_color(200, 200, 200)
             self.line(self.l_margin, self.get_y(), self.w - self.r_margin, self.get_y())
@@ -282,41 +289,37 @@ def _md_to_pdf(title: str, content_md: str) -> bytes:
     pdf.set_auto_page_break(auto=True, margin=15)
     pdf.add_page()
 
-    # Render markdown line by line with minimal formatting
+    def _cell(text: str, size: int = 10, bold: bool = False, indent: float = 0, h: float = 5):
+        try:
+            pdf.set_font("Helvetica", "B" if bold else "", size)
+            pdf.set_text_color(20 if bold else 30, 20 if bold else 30, 20 if bold else 30)
+            if indent:
+                pdf.set_x(pdf.l_margin + indent)
+            pdf.multi_cell(pdf.w - pdf.l_margin - pdf.r_margin - indent, h, _l1(text))
+        except Exception:
+            pass
+
     for raw_line in content_md.split("\n"):
         line = raw_line.rstrip()
         if line.startswith("# "):
-            pdf.set_font("Helvetica", "B", 14)
-            pdf.set_text_color(20, 20, 20)
-            pdf.multi_cell(0, 7, line[2:])
+            _cell(line[2:], size=14, bold=True, h=7)
             pdf.ln(1)
         elif line.startswith("## "):
-            pdf.set_font("Helvetica", "B", 12)
-            pdf.set_text_color(40, 40, 40)
-            pdf.multi_cell(0, 6, line[3:])
+            _cell(line[3:], size=12, bold=True, h=6)
             pdf.ln(1)
         elif line.startswith("### "):
-            pdf.set_font("Helvetica", "B", 11)
-            pdf.set_text_color(60, 60, 60)
-            pdf.multi_cell(0, 6, line[4:])
+            _cell(line[4:], size=11, bold=True, h=6)
         elif line.startswith(("- ", "* ", "+ ")):
-            pdf.set_font("Helvetica", "", 10)
-            pdf.set_text_color(30, 30, 30)
-            pdf.multi_cell(0, 5, "  •  " + line[2:])
+            _cell("•  " + line[2:], indent=5)
         elif re.match(r"^\d+\. ", line):
-            pdf.set_font("Helvetica", "", 10)
-            pdf.set_text_color(30, 30, 30)
-            pdf.multi_cell(0, 5, "  " + line)
-        elif line == "" or line == "---":
+            _cell(line, indent=5)
+        elif line in ("", "---"):
             pdf.ln(3)
         else:
-            # Strip inline markdown markers for bold/italic
             clean = re.sub(r"\*\*(.+?)\*\*", r"\1", line)
             clean = re.sub(r"\*(.+?)\*", r"\1", clean)
             clean = re.sub(r"`(.+?)`", r"\1", clean)
-            pdf.set_font("Helvetica", "", 10)
-            pdf.set_text_color(30, 30, 30)
-            pdf.multi_cell(0, 5, clean)
+            _cell(clean)
 
     return bytes(pdf.output())
 
@@ -775,8 +778,9 @@ def _process_analysis_job(job_id: int):
                     try:
                         pdf_bytes = _md_to_pdf(step_title, content)
                         attachments.append((f"{fname_base}.pdf", pdf_bytes, "application", "pdf"))
-                    except Exception:
-                        pass
+                    except Exception as pdf_err:
+                        import logging
+                        logging.warning(f"PDF-Generierung fehlgeschlagen für '{step_title}': {pdf_err}")
                 _send_email(email_to, subject, answer, attachments)
             except Exception as mail_err:
                 _job_set(job_id, error=f"[E-Mail-Fehler] {mail_err}")