docs(container): stop advertising private GHCR pulls

README.md

@@ -303,32 +303,23 @@ pip3 install -e ./
 
 SenseVoice can be built and run using Docker to simplify setup, ensure reproducibility, and support both CPU and GPU inference.
 
-### Official image
-
-The official `linux/amd64` runtime image is published to GitHub Container Registry. Model weights are not embedded in the image; they are downloaded into the mounted `/models` cache on first start.
-
-```bash
-docker pull ghcr.io/qwenaudio/sensevoice:latest
-docker run --rm --gpus all \
-  -p 50000:50000 \
-  -v sensevoice-models:/models \
-  ghcr.io/qwenaudio/sensevoice:latest
-```
-
-Open `http://127.0.0.1:50000/docs` after the container becomes healthy. Release tags and immutable `sha-<commit>` tags are available in the [container package](https://github.com/QwenAudio/SenseVoice/pkgs/container/sensevoice).
-
 ### Build with Docker
 ```bash
 docker build -t sensevoice .
 ```
 
+> The build workflow also publishes `ghcr.io/qwenaudio/sensevoice`, but the
+> package is currently private and anonymous pulls return HTTP 401. Use the
+> local build above until the [container package](https://github.com/QwenAudio/SenseVoice/pkgs/container/sensevoice)
+> is marked Public.
+
 ### Run (GPU – default)
 ```bash
 docker run --gpus all -p 50000:50000 sensevoice
 ```
 ### Run (CPU-only)
 ```bash
-docker run --rm -e SENSEVOICE_DEVICE=cpu -p 50000:50000 -v sensevoice-models:/models ghcr.io/qwenaudio/sensevoice:latest
+docker run --rm -e SENSEVOICE_DEVICE=cpu -p 50000:50000 -v sensevoice-models:/models sensevoice
 ```
 ### Docker Compose
 Docker Compose provides an easier way to run SenseVoice with persistent model caching, networking etc. 

README_zh.md

@@ -274,19 +274,29 @@ export SENSEVOICE_DEVICE=cuda:0
 fastapi run --port 50000
 ```
 
-### Docker 官方镜像
+### 使用 Docker 构建
 
-官方 `linux/amd64` 运行时镜像发布在 GitHub Container Registry。镜像不内置模型权重；首次启动会把权重下载到挂载的 `/models` 缓存。
+```bash
+docker build -t sensevoice .
+```
+
+> 构建工作流也会发布 `ghcr.io/qwenaudio/sensevoice`，但该容器包当前为
+> private，匿名拉取会返回 HTTP 401。在[容器包页面](https://github.com/QwenAudio/SenseVoice/pkgs/container/sensevoice)
+> 显示 Public 之前，请使用上面的本地构建。
+
+### 运行（GPU，默认）
+
+```bash
+docker run --rm --gpus all -p 50000:50000 -v sensevoice-models:/models sensevoice
+```
+
+### 运行（仅 CPU）
 
 ```bash
-docker pull ghcr.io/qwenaudio/sensevoice:latest
-docker run --rm --gpus all \
-  -p 50000:50000 \
-  -v sensevoice-models:/models \
-  ghcr.io/qwenaudio/sensevoice:latest
+docker run --rm -e SENSEVOICE_DEVICE=cpu -p 50000:50000 -v sensevoice-models:/models sensevoice
 ```
 
-容器健康后访问 `http://127.0.0.1:50000/docs`。版本 tag 和不可变的 `sha-<commit>` tag 见 [容器包页面](https://github.com/QwenAudio/SenseVoice/pkgs/container/sensevoice)。CPU 模式增加 `-e SENSEVOICE_DEVICE=cpu` 并移除 `--gpus all`。
+容器健康后访问 `http://127.0.0.1:50000/docs`。
 
 ## 微调
 

tests/test_container_contract.py

@@ -39,16 +39,25 @@ def test_container_workflow_builds_prs_and_only_pushes_trusted_refs(self):
             workflow,
         )
 
-    def test_readme_uses_public_ghcr_image_and_real_service_port(self):
+    def test_readme_uses_real_service_port_and_no_retired_registry(self):
         readme = (ROOT / "README.md").read_text(encoding="utf-8")
 
-        self.assertIn("ghcr.io/qwenaudio/sensevoice", readme.lower())
         self.assertIn("-p 50000:50000", readme)
         self.assertNotIn(
             "registry.cn-hangzhou.aliyuncs.com/funasr/sensevoice",
             readme,
         )
 
+    def test_readmes_do_not_offer_private_ghcr_image_as_anonymous_pull(self):
+        for name in ("README.md", "README_zh.md"):
+            readme = (ROOT / name).read_text(encoding="utf-8").lower()
+
+            self.assertNotIn(
+                "docker pull ghcr.io/qwenaudio/sensevoice:latest",
+                readme,
+            )
+            self.assertIn("docker build -t sensevoice .", readme)
+
 
 if __name__ == "__main__":
     unittest.main()