Update ADK doc according to issue #1521 - 3

docs/tools/third-party/crewai-and-langchain.md

@@ -0,0 +1,77 @@
+---
+title: CrewAI and LangChain integrations
+---
+
+# CrewAI and LangChain integrations
+
+The ADK provides wrapper classes that allow you to use third-party tools from [CrewAI](https://www.crewai.com/) and [LangChain](https://www.langchain.com/) in your ADK-based agents.
+
+## CrewAI integration
+
+To use a CrewAI tool, you need to wrap it with the `CrewaiTool` class. This adapter class converts the CrewAI tool into a format that is compatible with the ADK.
+
+### Example
+
+The following example shows how to use a CrewAI tool in an ADK agent.
+
+First, define your CrewAI tool:
+
+```python
+from crewai_tools import BaseTool
+
+class MyCrewAITool(BaseTool):
+    name: str = "My CrewAI Tool"
+    description: str = "A description of my CrewAI tool."
+
+    def _run(self, input: str) -> str:
+        return f"This is the output of my CrewAI tool with input: {input}"
+```
+
+Next, in your agent configuration, define a tool that uses the `CrewaiTool` wrapper to point to your CrewAI tool:
+
+```yaml
+tools:
+  - id: my_crewai_tool
+    tool_class: google.adk.integrations.crewai.CrewaiTool
+    tool_args:
+      name: "my_crewai_tool"
+      description: "A description of my CrewAI tool."
+      tool: "path.to.your.crewai_tool.MyCrewAITool"
+```
+
+Replace `"path.to.your.crewai_tool.MyCrewAITool"` with the actual import path to your tool.
+
+## LangChain integration
+
+To use a LangChain tool, you need to wrap it with the `LangchainTool` class. This adapter class converts the LangChain tool into a format that is compatible with the ADK.
+
+### Example
+
+The following example shows how to use a LangChain tool in an ADK agent.
+
+First, define your LangChain tool:
+
+```python
+from langchain_core.tools import BaseTool
+
+class MyLangChainTool(BaseTool):
+    name: str = "My LangChain Tool"
+    description: str = "A description of my LangChain tool."
+
+    def _run(self, input: str) -> str:
+        return f"This is the output of my LangChain tool with input: {input}"
+```
+
+Next, in your agent configuration, define a tool that uses the `LangchainTool` wrapper to point to your LangChain tool:
+
+```yaml
+tools:
+  - id: my_langchain_tool
+    tool_class: google.adk.integrations.langchain.LangchainTool
+    tool_args:
+      name: "my_langchain_tool"
+      description: "A description of my LangChain tool."
+      tool: "path.to.your.langchain_tool.MyLangChainTool"
+```
+
+Replace `"path.to.your.langchain_tool.MyLangChainTool"` with the actual import path to your tool.