diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..2f0aced
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,7 @@
+# Raw audio files are not tracked — only their transcripts (text) are.
+**/audio/
+*.mp3
+*.wav
+*.m4a
+*.ogg
+*.flac
diff --git a/A1.1/_template.md b/A1.1/_template.md
index 3ccf21d..61e7523 100644
--- a/A1.1/_template.md
+++ b/A1.1/_template.md
@@ -22,6 +22,16 @@ Source: [ ] Course book  [ ] Workbook
 
 - e.g. fill-in-the-blank conjugation, matching, listening comprehension
 
+## Audio / Listening
+
+Transcripts are ASR-generated (faster-whisper) unless noted otherwise —
+review for errors, especially names/numbers.
+
+### Track [X.X] — [short description, e.g. "Anna and Ben introduce themselves"]
+```
+[transcript text or path to transcripts/lektion-XX/trackname.txt]
+```
+
 ## Notes / gotchas
 
 - Anything tricky, false friends, exceptions worth flagging in future practice
diff --git a/README.md b/README.md
index ba1db9b..14b0f49 100644
--- a/README.md
+++ b/README.md
@@ -32,6 +32,25 @@ A1.2/                        ← added once I move on to the next book
 4. Update `INDEX.md` with a one-line entry for the lesson.
 5. Commit.
 
+## Audio (listening exercises)
+
+Course book audio has no printed transcript (the workbook's back section is
+the *Lösungsschlüssel* — answer key — not a transcript, so that doesn't help
+here). Raw audio files are **not** stored in this repo (see `.gitignore`);
+only their transcripts are, since that's what gives Claude context.
+
+Workflow:
+1. Get the audio files locally (publisher CD/app/download), any folder outside git.
+2. Run `scripts/transcribe.py` (uses `faster-whisper` on GPU) to transcribe them:
+   ```bash
+   pip install faster-whisper
+   python scripts/transcribe.py "<path-to-lesson-audio>" "A1.1/course-book/transcripts/lektion-XX"
+   ```
+3. Skim the `.txt` output for ASR mistakes (names, numbers, fast speech) and fix them.
+4. Reference/paste the transcript into the lesson's `.md` file under "Audio / Listening",
+   or just point at the `transcripts/lektion-XX/*.txt` path — either works, transcripts
+   are small text files so committing them is fine.
+
 ## Adding a new book (e.g. A1.2)
 
 Copy the `A1.1/` folder structure (minus content) into a new `A1.2/` folder
diff --git a/scripts/transcribe.py b/scripts/transcribe.py
new file mode 100644
index 0000000..52f35e6
--- /dev/null
+++ b/scripts/transcribe.py
@@ -0,0 +1,66 @@
+"""
+Transcribe course book listening-exercise audio to German text using
+faster-whisper (GPU-accelerated via CTranslate2).
+
+Setup (one time):
+    pip install faster-whisper
+
+Usage:
+    python transcribe.py <path-to-audio-folder> <path-to-output-folder>
+
+Example:
+    python transcribe.py "D:/MenschenA1.1/audio/Lektion01" "../A1.1/course-book/transcripts/lektion-01"
+
+Each audio file gets a matching .txt file with the same basename.
+Review the output — ASR on textbook dialogue audio is generally good but can
+mangle names, numbers, and fast/overlapping speech. Correct those before
+committing.
+"""
+
+import sys
+from pathlib import Path
+
+from faster_whisper import WhisperModel
+
+# "large-v3" is most accurate; drop to "medium" if VRAM is limited (~5GB vs ~10GB).
+MODEL_SIZE = "large-v3"
+DEVICE = "cuda"
+COMPUTE_TYPE = "float16"  # use "int8_float16" if you run out of VRAM
+
+AUDIO_EXTENSIONS = {".mp3", ".wav", ".m4a", ".ogg", ".flac"}
+
+
+def main() -> None:
+    if len(sys.argv) != 3:
+        print(__doc__)
+        sys.exit(1)
+
+    audio_dir = Path(sys.argv[1])
+    out_dir = Path(sys.argv[2])
+    out_dir.mkdir(parents=True, exist_ok=True)
+
+    model = WhisperModel(MODEL_SIZE, device=DEVICE, compute_type=COMPUTE_TYPE)
+
+    audio_files = sorted(
+        p for p in audio_dir.iterdir() if p.suffix.lower() in AUDIO_EXTENSIONS
+    )
+    if not audio_files:
+        print(f"No audio files found in {audio_dir}")
+        return
+
+    for audio_path in audio_files:
+        print(f"Transcribing {audio_path.name} ...")
+        segments, info = model.transcribe(str(audio_path), language="de", beam_size=5)
+
+        out_path = out_dir / (audio_path.stem + ".txt")
+        with out_path.open("w", encoding="utf-8") as f:
+            for segment in segments:
+                f.write(f"[{segment.start:6.1f}s] {segment.text.strip()}\n")
+
+        print(f"  -> {out_path}")
+
+    print("\nDone. Review transcripts for ASR errors before committing.")
+
+
+if __name__ == "__main__":
+    main()