diff --git a/README.md b/README.md
index f0f9e7a..dfe1011 100644
--- a/README.md
+++ b/README.md
@@ -189,7 +189,7 @@ audio:
 Without an `audio` config, voice messages from non-audio-capable models will display an error message to the user. For local transcription, you can run a whisper.cpp server:
 
 ```bash
-# Start whisper.cpp server with OpenAI-compatible endpoint
+# Option 1: Manual docker run
 docker run -d \
   --name whisper-server \
   -p 18801:8080 \
@@ -200,6 +200,9 @@ docker run -d \
   --convert \
   --language en \
   --inference-path /v1/audio/transcriptions
+
+# Option 2: Using docker-compose (uncomment whisper-server service in docker-compose.yml)
+# docker compose up -d
 ```
 
 ## Telegram Commands
diff --git a/docker-compose.yml b/docker-compose.yml
index d98b958..bafd2bd 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -30,5 +30,27 @@ services:
       start_period: 15s
       retries: 3
 
+  # Optional: Whisper server for audio transcription
+  # Uncomment and adjust as needed for voice message support
+  # whisper-server:
+  #   image: ghcr.io/ggml-org/whisper.cpp:main
+  #   container_name: whisper-server
+  #   restart: unless-stopped
+  #   ports:
+  #     - "18801:8080"
+  #   command: whisper-server
+  #     --model /app/models/ggml-base.en.bin
+  #     --host 0.0.0.0
+  #     --port 8080
+  #     --convert
+  #     --language en
+  #     --inference-path /v1/audio/transcriptions
+  #   healthcheck:
+  #     test: ["CMD-SHELL", "curl", "-f", "http://localhost:8080/"]
+  #     interval: 30s
+  #     timeout: 5s
+  #     start_period: 15s
+  #     retries: 3
+
 volumes:
   flynn-data: