Improve docker builds

Signed-off-by: Michael Yuan <[email protected]>

Author: juntao

docker/server-vad/Dockerfile

@@ -55,6 +55,7 @@ ENV LD_LIBRARY_PATH=/usr/local/lib/libtorch/lib:$LD_LIBRARY_PATH
 ENV LIBTORCH=/usr/local/lib/libtorch
 
 COPY config.toml .
+COPY hello.wav .
 
 RUN cat <<'EOF' > /usr/local/bin/start_servers.sh \
     && chmod +x /usr/local/bin/start_servers.sh

docker/server-vad/README.md

@@ -27,7 +27,11 @@ The VAD server listens on port `8000` internally. Choose one of the following so
 ## Build
 
 ```sh
-docker build -t secondstate/echokit:latest-server-vad -f docker/vad/Dockerfile .
+docker build \
+  --platform linux/amd64 \
+  --build-arg ECHOKIT_VERSION=0.1.0 \
+  -t secondstate/echokit:latest-server-vad \
+  .
 ```
 
 This Dockerfile always downloads the `linux-x86_64` release artifacts baked into the file. Update the URLs if you need to pin a different release.

docker/server-vad/config.toml

@@ -0,0 +1,29 @@
+addr = "0.0.0.0:8080"
+hello_wav = "hello.wav"
+
+[tts]
+platform = "Groq"
+model = "playai-tts"
+api_key = "gsk_ABCD"
+voice = "Fritz-PlayAI"
+
+[asr]
+url = "https://api.groq.com/openai/v1/audio/transcriptions"
+api_key = "gsk_ABCD"
+model = "whisper-large-v3"
+lang = "en"
+prompt = "Hello\n你好\n(noise)\n(bgm)\n(silence)\n"
+# Requires a local Silero VAD server: https://github.com/second-state/silero_vad_server
+vad_realtime_url = "ws://localhost:8000/v1/audio/realtime_vad"
+
+[llm]
+llm_chat_url = "https://api.groq.com/openai/v1/chat/completions"
+api_key = "gsk_ABCD"
+model = "openai/gpt-oss-20b"
+history = 5
+
+[[llm.sys_prompts]]
+role = "system"
+content = """
+You are a helpful assistant. Answer truthfully and concisely.
+"""

docker/server-vad/hello.wav

@@ -26,7 +26,9 @@ RUN set -eux; \
     test -n "$bin_path"; \
     install -m 0755 "$bin_path" /usr/local/bin/echokit_server; \
     rm -rf "$tmpdir" /tmp/echokit_server.tar.gz
-COPY config.toml config.toml
+
+COPY config.toml .
+COPY hello.wav .
 
 ENV RUST_LOG=info
 

docker/server/Dockerfile

@@ -21,7 +21,7 @@ Mount your `config.toml` directly into `/app/config.toml`. If you need to overri
 ## Build
 
 ```sh
-docker build -t secondstate/echokit:latest-server -f docker/server/Dockerfile .
+docker build -t secondstate/echokit:latest-server .
 ```
 
 The build automatically selects the correct release artifact for your build architecture using the BuildKit-provided `TARGETPLATFORM`/`TARGETARCH` arguments. Override the downloaded release by supplying `--build-arg ECHOKIT_VERSION=<version>` (for example `0.1.1`) if you want a different tag.
@@ -33,10 +33,9 @@ Use Buildx to produce and publish a multi-arch manifest in one command. BuildKit
 ```sh
 docker buildx build \
   --platform linux/amd64,linux/arm64 \
-  --push \
   --build-arg ECHOKIT_VERSION=0.1.0 \
   --tag secondstate/echokit:latest-server \
-  -f docker/server/Dockerfile .
+  .
 ```
 
 Adjust `ECHOKIT_VERSION` as needed; omit the flag to fall back to the default version baked into the Dockerfile.

docker/server/README.md

@@ -0,0 +1,27 @@
+addr = "0.0.0.0:8080"
+hello_wav = "hello.wav"
+
+[tts]
+platform = "Groq"
+model = "playai-tts"
+api_key = "gsk_ABCD"
+voice = "Fritz-PlayAI"
+
+[asr]
+url = "https://api.groq.com/openai/v1/audio/transcriptions"
+api_key = "gsk_ABCD"
+model = "whisper-large-v3"
+lang = "en"
+prompt = "Hello\n你好\n(noise)\n(bgm)\n(silence)\n"
+
+[llm]
+llm_chat_url = "https://api.groq.com/openai/v1/chat/completions"
+api_key = "gsk_ABCD"
+model = "openai/gpt-oss-20b"
+history = 5
+
+[[llm.sys_prompts]]
+role = "system"
+content = """
+You are a helpful assistant. Answer truthfully and concisely.
+"""