diff --git a/docs/OPCUA_AUTODISCOVERY.md b/docs/OPCUA_AUTODISCOVERY.md
new file mode 100644
index 0000000..79b2019
--- /dev/null
+++ b/docs/OPCUA_AUTODISCOVERY.md
@@ -0,0 +1,69 @@
+# OPCUA Auto Discovery with SensApp.
+
+Unlike MQTT, OPCUA does not have a subscription mechanism allowing wildcards. One must specify all the nodes to monitor on the OPCUA server.
+
+This is not always very convenient, so to streamline the deployment of SensApp, SensApp supports the auto-discovery of OPCUA variable nodes.
+
+SensApp will browse the OPCUA server and automatically monitors the variables and collect the data from the OPCUA server.
+
+## Configuration
+
+```toml
+[opcua.subscriptions.autodiscovery]
+
+# Enable or disable the auto-discovery of OPCUA nodes. Enabled by default if the configuration is present.
+enabled = true
+
+# Start node for the auto-discovery. It will be the root node by default.
+# Some server contain a LOT of nodes under the root node, so you may prefer to start
+# from another node. It uses the same
+# syntax as the identifiers in the OPCUA subscrptions.
+start_node = "DataBlocksGlobal"
+
+# List of node identifiers that are excluded from the auto-discovery.
+# They will not be monitored or navigated through.
+# By default, no nodes are excluded.
+excluded_nodes = [
+    "DataBlocksGlobal/SomeNode",
+    "DataBlocksGlobal/AnotherNode"
+]
+
+# Maintaining a list of excluded nodes can be cumbersome too. You can use a regular expression to exclude nodes based on their
+# browse name.
+# By default, no nodes are excluded.
+node_browse_name_exclude_regex = "^(Icon|History)$"
+
+# Sometimes you will discover many variables that you do not want to monitor.
+# You can use a regular expression to include only the variables that have an identifier that match the regular expression.
+# If it's not a string, it uses the string representation of the identifier.
+# By default, all the variables are included.
+variable_identifier_include_regex = ".*Temperature.*"
+
+# The auto-discovery starts from the namespace fo the subscription or the root node, but it can discover nodes in other namespaces if the OPCUA server is configured with links accross namespaces.
+# This may not be desirable, so you can ignore the nodes from other namespaces.
+# Enabled by default.
+discover_accross_namespaces = true
+
+# Some OPCUA servers will organise some variables under a parent node
+# that is also a variable, of type object.
+# This is nice, but if you don't want to monitor the parent node as well as the children, you can use this option.
+# It is enabled by default.
+skip_variables_with_children = true
+
+# Maximum discovery depth.
+# 32 by default
+max_depth = 32
+
+# Maximum number of nodes to discover.
+# 1024 by default
+max_nodes = 1024
+
+```
+
+## In Production
+
+In production, you may not want to rely on the auto-discovery. Also, the auto-discovery is done only once at the start of the subscription. It will not monitor for new nodes that are added to the OPCUA server, or nodes that are removed.
+
+Moreover, the auto-discovery is not very fast, and it may take a little while before the data is being collected.
+
+Still, this may work well for your use case. As a reminder, this software is provided as-is without any warranty.