TexasInstruments · jeevantelukula · Dec 17, 2025 · Nov 26, 2025
@@ -100,6 +100,7 @@ linux/Foundational_Components/Virtualization/Docker
 linux/Foundational_Components_Machine_Learning
 linux/Foundational_Components/Machine_Learning/armnn
 linux/Foundational_Components/Machine_Learning/tflite
+linux/Foundational_Components/Machine_Learning/nnstreamer
 
 linux/Foundational_Components/Graphics/index
 linux/Foundational_Components/Graphics/Common/Display
@@ -113,6 +114,9 @@ linux/Foundational_Components/Graphics/SGX/Overview
 linux/Foundational_Components/Graphics/SGX/SGX_Debug_Info
 linux/Foundational_Components/Graphics/SGX/Build_Guide
 
+linux/Demo_User_Guides/index_Demos
+linux/Demo_User_Guides/Webserver_Demo_User_Guide
+
 linux/Examples_and_Demos
 linux/Examples_and_Demos_Application_Demos
 linux/Examples_and_Demos/Application_Demos/QT_Thermostat_HMI_Demo

@@ -16,6 +16,7 @@ Processor SDK Linux Software Developer's Guide
    /linux/Overview
    Release_Specific
    /linux/Foundational_Components
+   /linux/Demo_User_Guides/index_Demos
    /linux/Industrial_Protocols
    /linux/Examples_and_Demos
    /linux/How_to_Guides

@@ -0,0 +1,166 @@
+.. _Webserver-Demo-User-Guide-label:
+
+############################
+Web server demo - User Guide
+############################
+
+********
+Overview
+********
+
+This document describes the Out of Box Web Server Demo Application, which the Linux
+SDK delivers with the |__SDK_FULL_NAME__| |__PART_FAMILY_NAME__|, applicable to boards
+such as the `BEAGL-BONE-GRN-ECO <https://www.ti.com/tool/BEAGL-BONE-GRN-ECO>`__ and
+`TMDXEVM3358 <https://www.ti.com/tool/TMDXEVM3358>`__. The main purpose of this demo is to
+showcase a glimpse of the ARM analytics AI stack integrated into the filesystem,
+providing an out-of-box experience even users do not connect a display to the board.
+
+The demo showcases a web server running on an AM335x based board, providing a web interface with three key sections:
+
+*  **Audio Classification Demo**: Users can connect a USB audio capture device to run real-time audio classification, displaying classification statistics.
+*  **Live CPU Performance Metrics**: Provides a real-time CPU usage indicator, historical usage data for the last 5 minutes, and detailed CPU information.
+*  **Documentation Links**: Offers convenient access to relevant documentation.
+
+This document provides all necessary equipment requirements and instructions.
+
+**********************
+Hardware Prerequisites
+**********************
+
+-  TI AM335x based board (e.g., `BEAGL-BONE-GRN-ECO <https://www.ti.com/tool/BEAGL-BONE-GRN-ECO>`__ and `TMDXEVM3358 <https://www.ti.com/tool/TMDXEVM3358>`__)
+   (Note: For the Beaglebone Green Eco, a display is available via an `HDMI cape <https://wiki.seeedstudio.com/BeagleBone_Green_HDMI_Cape>`__)
+-  PC (Windows or Linux)
+-  Ethernet cables
+-  Ethernet switch or ethernet router with DHCP service
+-  SD card (minimum 16GB)
+-  Audio Capture Device
+
+***********
+Get Started
+***********
+
+#.  Flash an SD card with the :file:`tisdk-default-image`. User can download the :file:`tisdk-default-image` wic
+    image from |__SDK_DOWNLOAD_URL__|. Please follow the instructions from here to :ref:`Flash an SD card <processor-sdk-linux-create-sd-card>`.
+
+#.  Insert the SD card into the AM335x based board and set it to boot via SD card Boot mode.
+
+#.  Connect an ethernet cable from your ethernet switch or router to the
+    AM335x based board.
+
+#.  Connect your PC to the same ethernet switch or router.
+
+#.  Connect the UART to the PC's USB port.
+
+#.  Open a terminal program (like TeraTerm or minicom) and connect to the
+    serial port. Set the port to 115200bps, 8 bit, no parity, 1 stop bit, no flow control.
+
+#.  Power on the AM335x based board.
+
+#.  After the Linux boot completes, login as "root". Use the :command:`ifconfig` command
+    to find out the IP address of the device.
+
+#. On the host PC, open a Internet Browser and enter in the following: ``http://<IP_ADDRESS_OF_AM335x_Board>:3000``
+
+#. The following web page will be displayed, the home page shows as below.
+
+.. image:: /images/Webserver_home_page.PNG
+   :alt: Webserver Demo Page
+   :width: 75%
+
+*************************
+Audio Classification Demo
+*************************
+
+The web interface provides an audio classification demo that uses the integrated ARM analytics AI stack.
+
+.. image:: /images/Webserver_audio_classification.PNG
+   :alt: Audio Classification Demo
+   :width: 75%
+
+To use the demo:
+
+#.  Connect a USB audio device (e.g., a USB headset) to the AM335x based board.
+#.  Click the "Refresh Devices" button to detect the connected audio capture device.
+#.  Select the desired audio capture device from the list of available devices.
+#.  Click the "Start Classification" button to begin real-time audio classification.
+#.  The "Live Classification" section will display the current classification, total classifications, unique classes, session time, and updates per minute.
+#.  A "Classification History" log is also available, showing a timestamped record of classifications.
+#.  Click the "Stop Classification" button to end the demo.
+
+**Technical Details**
+
+The audio classification demo leverages :ref:`NNStreamer <nnstreamer-label>`, a GStreamer-based neural network framework, to create and run the processing pipeline.
+
+*  **Model:** The demo uses the `YAMNet <https://www.kaggle.com/models/google/yamnet/tfLite>`__ sound classification model.
+*  **Deep Learning Runtime:** The model is executed using the :ref:`TensorFlow Lite <tflite-label>` runtime.
+
+**GStreamer Pipeline**
+
+The following GStreamer pipeline is used to process the audio from the microphone, run the inference, and decode the results:
+
+.. code-block:: console
+
+   gst-launch-1.0 alsasrc device=<device> ! \
+       audioconvert ! \
+       audio/x-raw,format=S16LE,channels=1,rate=16000,layout=interleaved ! \
+       tensor_converter frames-per-tensor=3900 ! \
+       tensor_aggregator frames-in=3900 frames-out=15600 frames-flush=3900 frames-dim=1 ! \
+       tensor_transform mode=arithmetic option=typecast:float32,add:0.5,div:32767.5 ! \
+       tensor_transform mode=transpose option=1:0:2:3 ! \
+       queue leaky=2 max-size-buffers=10 ! \
+       tensor_filter framework=tensorflow2-lite model=/usr/share/oob-demo-assets/models/yamnet_audio_classification.tflite custom=Delegate:XNNPACK,NumThreads:2 ! \
+       tensor_decoder mode=image_labeling option1=/usr/share/oob-demo-assets/labels/yamnet_label_list.txt ! \
+       filesink buffer-mode=2 location=<fifo_path>
+
+****************************
+Live CPU Performance Metrics
+****************************
+
+The web interface also provides a live view of the CPU performance metrics.
+
+.. image:: /images/Webserver_CPU_Performance.PNG
+   :alt: CPU Performance Metrics
+   :width: 75%
+
+This section includes:
+
+*  **CPU Usage**: A real-time gauge showing the current CPU utilization.
+*  **CPU History**: A graph displaying the CPU usage history over the last 5 minutes.
+*  **Statistics**: Average and maximum CPU usage calculated for the data from CPU History data.
+
+*********************
+Software Architecture
+*********************
+
+The demo consists of three main components: a web interface, a backend web server, and a set of Linux applications.
+
+*  **Web Interface (GUI):** A dynamic HTML page with JavaScript that runs in the user's browser. It provides the user interface for the audio classification demo and CPU performance metrics, making asynchronous requests to the web server for data updates.
+
+*  **Web Server:** A lightweight HTTP server (e.g., Node.js) running on the AM335x's Arm A8 core. It serves the static web page and provides a simple REST API for the frontend to interact with the underlying Linux applications.
+
+*  **Linux Applications:** A set of background applications running on the AM335x, with the following key functions:
+
+    *  A C application that reads real-time system information (like CPU stats from :file:`/proc/stat`) and makes it available to the web server.
+    *  The audio classification demo, which is a GStreamer pipeline that leverages the NNStreamer framework. This pipeline reads from a USB audio device, processes the audio with the YAMNet model using the TensorFlow Lite runtime, and sends the classification results to the web server.
+
+*******************
+Directory Structure
+*******************
+
+Yocto recipe for building this demo can be found at `github:webserver-oob_git.bb <https://github.com/TexasInstruments/meta-tisdk/blob/scarthgap/meta-ti-foundational/recipes-demos/webserver-oob/webserver-oob_git.bb>`__.
+
+The source code for the demo can be found at `github/TexasInstruments/webserver-oob-demo <https://github.com/TexasInstruments/webserver-oob-demo>`__.
+For instructions on building the utilities see the `README <https://github.com/TexasInstruments/webserver-oob-demo/blob/main/README.md>`__ file.
+
+.. list-table:: Directory Structure
+   :widths: 20 30
+   :header-rows: 1
+
+   * - Directory Name
+     - Description
+   * - :file:`webserver_app/app`
+     - GUI code (HTML, CSS, JavaScript)
+   * - :file:`webserver_app/linux_app`
+     - Linux application code (C code)
+   * - :file:`webserver_app/webserver`
+     - Web server code (Node.js)
@@ -14,6 +14,7 @@ The SDK supports the following Out-Of-Box demo-applications
    Seva_Store
    Chromium_Browser
    Benchmark_Demo_User_Guide
+   Webserver_Demo_User_Guide
    Display_Cluster_User_Guide
    TI_LVGL_Demo_User_Guide
    AM62D_Dsp_Offload_User_Guide.rst

@@ -1,3 +1,5 @@
+.. _nnstreamer-label:
+
 NNStreamer
 ==========
 
@@ -37,7 +39,7 @@ Run the included unit tests:
 To list available elements:
 
 .. code-block:: console
-      
+
     root@am62xx-evm:~# gst-inspect-1.0 | grep nnstreamer
     nnstreamer:  tensor_aggregator: TensorAggregator
     nnstreamer:  tensor_converter: TensorConverter

@@ -1,3 +1,5 @@
+.. _tflite-label:
+
 ########################
 TensorFlow Lite (LiteRT)
 ########################
@@ -176,22 +178,22 @@ The following performance numbers are captured with :command:`benchmark_model` o
 Based on the above data, using the XNNPACK delegate significantly improves inference times across all SoCs, though it generally increases initialization time and overall memory footprint.
 
 .. note::
-   
+
    The performance numbers mentioned above were recorded after stopping the out-of-box (OOB) demos included in the TI SDK.
-   
+
 ********************
 Example Applications
 ********************
- 
-|__SDK_FULL_NAME__| has integrated opensource components like NNStreamer which can be used for neural network inferencing using the sample tflite models under :file:`/usr/share/oob-demo-assets/models/` 
+
+|__SDK_FULL_NAME__| has integrated opensource components like NNStreamer which can be used for neural network inferencing using the sample tflite models under :file:`/usr/share/oob-demo-assets/models/`
 Checkout the Object Detection usecase under :ref:`TI Apps Launcher - User Guide <TI-Apps-Launcher-User-Guide-label>`
 
 Alternatively, if a display is connected, you can run the Object Detection pipeline using this command,
 
 .. ifconfig:: CONFIG_part_variant in ('AM62X', 'AM62LX')
 
    .. code-block:: console
-   
+
       gst-launch-1.0 multifilesrc location=/usr/share/oob-demo-assets/videos/oob-gui-video-objects.h264 loop=true ! \
       h264parse ! avdec_h264 ! \
       tee name=tee_split0 \
@@ -220,7 +222,7 @@ Alternatively, if a display is connected, you can run the Object Detection pipel
 .. ifconfig:: CONFIG_part_variant in ('AM62PX')
 
    .. code-block:: console
-   
+
       gst-launch-1.0 multifilesrc location=/usr/share/oob-demo-assets/videos/oob-gui-video-objects.h264 loop=true caps=video/x-h264,width=1280,height=720,framerate=1/1 ! \
       h264parse ! v4l2h264dec capture-io-mode=4 ! \
       tee name=tee_split0 \
@@ -247,5 +249,5 @@ Alternatively, if a display is connected, you can run the Object Detection pipel
    The above GStreamer pipeline reads an H.264 video file, decodes it, and processes it for object detection using a TensorFlow Lite model, displaying bounding boxes around detected objects. The processed video is then composited and rendered on the screen using the ``kmssink`` element.
 
 .. attention::
-   
+
    The Example Applications section is not applicable for AM64x