How to install and use the xpander AI SDK
Overview
This library provides a Python client to interact with xpander AI's platform and retrieve tools, invoke tools and use local tools for function calling against various of LLM Providers such as Nvidia NIM, OpenAI and Amazon Bedrock
The library runs as a NodeJS app under the hood and compiled safely using Projen In order to run our SDK smoothly you must have NodeJS installed.
Installing the SDK
pip install https://assets.xpanderai.io/xpander-sdk.tar.gz
conda install -c xpanderai xpander-sdk
npm install https://assets.xpanderai.io/xpander-sdk.tar.gz
Install-Package Xpander.Sdk -Source https://assets.xpanderai.io/nuget
<dependency>
<groupId>ai.xpander</groupId>
<artifactId>xpander-sdk</artifactId>
<version>1.0.0</version>
</dependency>
<repositories>
<repository>
<id>xpander-repo</id>
<url>https://assets.xpanderai.io/maven</url>
</repository>
</repositories>
Retrieving the SDK's agent Key and agent URL
In the xpander AI's platform, go to AI Apps > New AI App > choose your Agent.
Copy the agent key and agent url from the platform.
Initializing the SDK
from xpander_sdk import XpanderClient, LLMProvider,OpenAISupportedModels
xpander_client = XpanderClient(
agent_key="XPANDER_AI_AGENT_KEY",
agent_url="XPANDER_AI_AGENT_URL",
llm_provider=LLMProvider.OPEN_AI # or use any other provider you want (NVIDIA_NIM, AMAZON_BEDROCK)
)
import { XpanderClient } from 'xpander-sdk'
xpanderClient = new XpanderClient(
"XPANDER_AI_AGENT_KEY",
"XPANDER_AI_AGENT_URL",
LLMProvider.OPEN_AI /* or use any other provider you want (NVIDIA_NIM, AMAZON_BEDROCK)*/
)
using Xpander.AI.Sdk;
XpanderClient xpanderClient = new XpanderClient(
"XPANDER_AI_AGENT_KEY",
"XPANDER_AI_AGENT_URL",
LLMProvider.OPEN_AI);
import ai.xpander.sdk.XpanderClient;
XpanderClient xpanderClient = new XpanderClient(
"XPANDER_AI_AGENT_KEY",
"XPANDER_AI_AGENT_URL",
LLMProvider.OPEN_AI /* or use any other provider you want (NVIDIA_NIM, AMAZON_BEDROCK)*/
);
SDK Methods
Get tools
our SDK allows you to retrieve tools connected in the xpander ai's platform and use it with ease.
when creating an AI App you need to choose connectors that you want to communicate with.
Request
# API Reference https://docs.xpander.ai/reference/python-sdk#xpander-sdk.XpanderClient.tools
tools = xpander_client.tools()
// API Reference https://docs.xpander.ai/reference/nodejs-sdk#xpander-sdk.XpanderClient.tools
tools = xpanderClient.tools()
// API Reference https://docs.xpander.ai/reference/net-sdk#xpander-sdk.XpanderClient.tools
IOpenAIToolOutput[] tools = xpanderClient.Tools();
// API Reference https://docs.xpander.ai/reference/java-sdk#xpander-sdk.XpanderClient.tools
IOpenAIToolOutput[] tools = xpanderClient.tools();
Response (List[IOpenAIToolOutput] | List[IBedrockToolOutput])
[
{
"type": "function",
"function": {
"name": "Conduit-article-management-createNewArticleWithTags",
"description": "creates new article system. Article body, description, title mandatory . run Conduit-article-management-retrieveArticlesWithOptionalFilters . IMPORTANT! make sure to use body_params, query_params, path_params. these are crucial for ensuring function calling works!",
"parameters": {
"type": "object",
"properties": {
"query_params": {
"type": "object",
"properties": {},
"required": []
},
"path_params": {
"type": "object",
"properties": {},
"required": []
},
"body_params": {
"type": "object",
"properties": {
"article": {
"type": "object",
"description": "article",
"properties": {
"body": {
"type": "string"
},
"description": {
"type": "string"
},
"tagList": {
"type": "array",
"items": {
"type": "string"
}
},
"title": {
"type": "string"
}
}
}
},
"required": [
"article"
]
}
},
"required": [
"query_params",
"path_params",
"body_params"
]
}
}
}
]
[
{
"type": "function",
"function": {
"name": "Conduit-article-management-createNewArticleWithTags",
"description": "creates new article system. Requires authentication. Article body, description, title r mandatory . run Conduit-article-management-retrieveArticlesWithOptionalFilters . IMPORTANT! make sure to use body_params, query_params, path_params. these are crucial for ensuring function calling works!",
"parameters": {
"type": "object",
"properties": {
"query_params": {
"type": "object",
"properties": {},
"required": []
},
"path_params": {
"type": "object",
"properties": {},
"required": []
},
"body_params": {
"type": "object",
"properties": {
"article": {
"type": "object",
"description": "article",
"properties": {
"body": {
"type": "string"
},
"description": {
"type": "string"
},
"tagList": {
"type": "array",
"items": {
"type": "string"
}
},
"title": {
"type": "string"
}
}
}
},
"required": [
"article"
]
}
},
"required": [
"query_params",
"path_params",
"body_params"
]
}
}
}
]
[
{
"toolSpec": {
"name": "Conduit_article_management_createNewArticleWithTags",
"description": "creates new article system. Requires authentication. Article body, description, title r mandatory . run Conduit-article-management-retrieveArticlesWithOptionalFilters . IMPORTANT! make sure to use body_params, query_params, path_params. these are crucial for ensuring function calling works!",
"inputSchema": {
"json": {
"type": "object",
"properties": {
"query_params": {
"type": "object",
"properties": {},
"required": []
},
"path_params": {
"type": "object",
"properties": {},
"required": []
},
"body_params": {
"type": "object",
"properties": {
"article": {
"type": "object",
"description": "article",
"properties": {
"body": {
"type": "string"
},
"description": {
"type": "string"
},
"tagList": {
"type": "array",
"items": {
"type": "string"
}
},
"title": {
"type": "string"
}
}
}
},
"required": [
"article"
]
}
},
"required": [
"query_params",
"path_params",
"body_params"
]
}
}
}
}
]
Using local functions
Our SDK allows you to retrieve tools connected in the xpander ai's platform and use it with ease.
sometime you might need your own local functions, here's the way to do it.
With local function support, you can create AI Agents utilizing the xpander connector library. When an LLM needs to invoke a locally stored function, such as for mathematical calculations, our SDK will return an object ready for local execution.
Request (List[ILocalTool])
# API Reference https://docs.xpander.ai/reference/python-sdk#xpander-sdk.XpanderClient.tools
local_tools = [
{
'type': 'function',
'function': {
'name': 'greet-by-name',
'description': 'Returns a greet to a user with name',
'parameters': {
'type': 'object',
'properties': {
'userName': {
'type': 'string',
'description': "The user's name",
},
},
'required': ['userName'],
'additionalProperties': False,
},
},
},
]
xpander_client.add_local_tools(local_tools)
tools = xpander_client.tools()
// API Reference https://docs.xpander.ai/reference/nodejs-sdk#xpander-sdk.XpanderClient.tools
tools = xpanderClient.tools()
// API Reference https://docs.xpander.ai/reference/net-sdk#xpander-sdk.XpanderClient.tools
IOpenAIToolOutput[] tools = xpanderClient.Tools();
// API Reference https://docs.xpander.ai/reference/java-sdk#xpander-sdk.XpanderClient.tools
IOpenAIToolOutput[] tools = xpanderClient.tools();
Response (List[IOpenAIToolOutput] | List[IBedrockToolOutput])
[
{
"type": "function",
"function": {
"name": "Conduit-article-management-createNewArticleWithTags",
"description": "creates new article system. Requires authentication. Article body, description, title r mandatory . run Conduit-article-management-retrieveArticlesWithOptionalFilters . IMPORTANT! make sure to use body_params, query_params, path_params. these are crucial for ensuring function calling works!",
"parameters": {
"type": "object",
"properties": {
"query_params": {
"type": "object",
"properties": {},
"required": []
},
"path_params": {
"type": "object",
"properties": {},
"required": []
},
"body_params": {
"type": "object",
"properties": {
"article": {
"type": "object",
"description": "article",
"properties": {
"body": {
"type": "string"
},
"description": {
"type": "string"
},
"tagList": {
"type": "array",
"items": {
"type": "string"
}
},
"title": {
"type": "string"
}
}
}
},
"required": [
"article"
]
}
},
"required": [
"query_params",
"path_params",
"body_params"
]
}
}
}
]
[
{
"type": "function",
"function": {
"name": "Conduit-article-management-createNewArticleWithTags",
"description": "creates new article system. Requires authentication. Article body, description, title r mandatory . run Conduit-article-management-retrieveArticlesWithOptionalFilters . IMPORTANT! make sure to use body_params, query_params, path_params. these are crucial for ensuring function calling works!",
"parameters": {
"type": "object",
"properties": {
"query_params": {
"type": "object",
"properties": {},
"required": []
},
"path_params": {
"type": "object",
"properties": {},
"required": []
},
"body_params": {
"type": "object",
"properties": {
"article": {
"type": "object",
"description": "article",
"properties": {
"body": {
"type": "string"
},
"description": {
"type": "string"
},
"tagList": {
"type": "array",
"items": {
"type": "string"
}
},
"title": {
"type": "string"
}
}
}
},
"required": [
"article"
]
}
},
"required": [
"query_params",
"path_params",
"body_params"
]
}
}
}
]
[
{
"toolSpec": {
"name": "Conduit_article_management_createNewArticleWithTags",
"description": "creates new article system. Requires authentication. Article body, description, title r mandatory . run Conduit-article-management-retrieveArticlesWithOptionalFilters . IMPORTANT! make sure to use body_params, query_params, path_params. these are crucial for ensuring function calling works!",
"inputSchema": {
"json": {
"type": "object",
"properties": {
"query_params": {
"type": "object",
"properties": {},
"required": []
},
"path_params": {
"type": "object",
"properties": {},
"required": []
},
"body_params": {
"type": "object",
"properties": {
"article": {
"type": "object",
"description": "article",
"properties": {
"body": {
"type": "string"
},
"description": {
"type": "string"
},
"tagList": {
"type": "array",
"items": {
"type": "string"
}
},
"title": {
"type": "string"
}
}
}
},
"required": [
"article"
]
}
},
"required": [
"query_params",
"path_params",
"body_params"
]
}
}
}
}
]
Invoke tools
when an LLM choose a tool, the tool should be invoked.
this is how you do it.
Request
xpander_client.xpander_tool_call(tool_selector_response=dict(llm_response))
const toolResponse = xpanderClient.xpanderToolCall(llmResponse);
ToolResponse[] toolResponses = xpanderClient.XpanderToolCall(llmResponse);
ToolResponse[] toolResponses = xpanderClient.xpanderToolCall(llmResponse);
Response (List[ToolResponse])
[
{
"name": "Conduit-article-management-getAllTagsForArticles",
"payload_property1": null,
"payload_property2": null,
"raw_response": "{\"tags\":[\"Pet Adventures\",\"BYD\",\"Xpander\",\"dogs\",\"nutrition\",\"Culture\",\"test-oren\",\"test\",\"workshop\",\"xpander.ai\",\"xpanderAI Agent\",\"Lucky\",\"Generative AI\",\"culinary\",\"quantum computing\",\"Rosy\",\"workflow automation\",\"xpanderAI\",\"pet\",\"Issues\",\"example\",\"gig economy\",\"business\",\"??????\",\"electric vehicle\",\"automotive\",\"AWS\",\"ert\",\"Leo\",\"AI Integration\",\"climate change\",\"animals\",\"urban planning\",\"pets\",\"Freddy\",\"fruits\",\"GitHub\",\"augmented reality\",\"conference\",\"Data Pipelines\",\"remote work\",\"history\",\"Testing\",\"animal\",\"Jewel\",\"DriveNets\",\"Nature\",\"Customer Service\",\"Bird\",\"tag1,tag2\",\"Machine Learning\",\"tag1\",\"Collaboration\",\"Lizard\",\"success\",\"apples\",\"adventures\",\"biotechnology\",\"Data Science\",\"gaming\",\"story\",\"computers\",\"sustainable agriculture\",\"AI\",\"Customer Support\",\"Network Management\",\"mental health\",\"5G\",\"Data Security\",\"funny stories\",\"sustainable fashion\",\"smart cities\",\"technology\",\"Ecology\",\"health\",\"adventure\",\"App4\",\"Fun\",\"Summary\",\"space exploration\",\"\\\"training\\\"]\",\"artificial intelligence\"]}",
"response_message": "{\"tags\":[\"Pet Adventures\",\"BYD\",\"Xpander\",\"dogs\",\"nutrition\",\"Culture\",\"test-oren\",\"test\",\"workshop\",\"xpander.ai\",\"xpanderAI Agent\",\"Lucky\",\"Generative AI\",\"culinary\",\"quantum computing\",\"Rosy\",\"workflow automation\",\"xpanderAI\",\"pet\",\"Issues\",\"example\",\"gig economy\",\"business\",\"??????\",\"electric vehicle\",\"automotive\",\"AWS\",\"ert\",\"Leo\",\"AI Integration\",\"climate change\",\"animals\",\"urban planning\",\"pets\",\"Freddy\",\"fruits\",\"GitHub\",\"augmented reality\",\"conference\",\"Data Pipelines\",\"remote work\",\"history\",\"Testing\",\"animal\",\"Jewel\",\"DriveNets\",\"Nature\",\"Customer Service\",\"Bird\",\"tag1,tag2\",\"Machine Learning\",\"tag1\",\"Collaboration\",\"Lizard\",\"success\",\"apples\",\"adventures\",\"biotechnology\",\"Data Science\",\"gaming\",\"story\",\"computers\",\"sustainable agriculture\",\"AI\",\"Customer Support\",\"Network Management\",\"mental health\",\"5G\",\"Data Security\",\"funny stories\",\"sustainable fashion\",\"smart cities\",\"technology\",\"Ecology\",\"health\",\"adventure\",\"App4\",\"Fun\",\"Summary\",\"space exploration\",\"\\\"training\\\"]\",\"artificial intelligence\"]}",
"role": "tool",
"tool_call_id": "tooluse_8ZEYkp2qS_aBpx5cR9XDzA"
}
]
[
{
"name": "Conduit-article-management-getAllTagsForArticles",
"payload_property1": null,
"payload_property2": null,
"raw_response": "{\"tags\":[\"Pet Adventures\",\"BYD\",\"Xpander\",\"dogs\",\"nutrition\",\"Culture\",\"test-oren\",\"test\",\"workshop\",\"xpander.ai\",\"xpanderAI Agent\",\"Lucky\",\"Generative AI\",\"culinary\",\"quantum computing\",\"Rosy\",\"workflow automation\",\"xpanderAI\",\"pet\",\"Issues\",\"example\",\"gig economy\",\"business\",\"??????\",\"electric vehicle\",\"automotive\",\"AWS\",\"ert\",\"Leo\",\"AI Integration\",\"climate change\",\"animals\",\"urban planning\",\"pets\",\"Freddy\",\"fruits\",\"GitHub\",\"augmented reality\",\"conference\",\"Data Pipelines\",\"remote work\",\"history\",\"Testing\",\"animal\",\"Jewel\",\"DriveNets\",\"Nature\",\"Customer Service\",\"Bird\",\"tag1,tag2\",\"Machine Learning\",\"tag1\",\"Collaboration\",\"Lizard\",\"success\",\"apples\",\"adventures\",\"biotechnology\",\"Data Science\",\"gaming\",\"story\",\"computers\",\"sustainable agriculture\",\"AI\",\"Customer Support\",\"Network Management\",\"mental health\",\"5G\",\"Data Security\",\"funny stories\",\"sustainable fashion\",\"smart cities\",\"technology\",\"Ecology\",\"health\",\"adventure\",\"App4\",\"Fun\",\"Summary\",\"space exploration\",\"\\\"training\\\"]\",\"artificial intelligence\"]}",
"response_message": "{\"tags\":[\"Pet Adventures\",\"BYD\",\"Xpander\",\"dogs\",\"nutrition\",\"Culture\",\"test-oren\",\"test\",\"workshop\",\"xpander.ai\",\"xpanderAI Agent\",\"Lucky\",\"Generative AI\",\"culinary\",\"quantum computing\",\"Rosy\",\"workflow automation\",\"xpanderAI\",\"pet\",\"Issues\",\"example\",\"gig economy\",\"business\",\"??????\",\"electric vehicle\",\"automotive\",\"AWS\",\"ert\",\"Leo\",\"AI Integration\",\"climate change\",\"animals\",\"urban planning\",\"pets\",\"Freddy\",\"fruits\",\"GitHub\",\"augmented reality\",\"conference\",\"Data Pipelines\",\"remote work\",\"history\",\"Testing\",\"animal\",\"Jewel\",\"DriveNets\",\"Nature\",\"Customer Service\",\"Bird\",\"tag1,tag2\",\"Machine Learning\",\"tag1\",\"Collaboration\",\"Lizard\",\"success\",\"apples\",\"adventures\",\"biotechnology\",\"Data Science\",\"gaming\",\"story\",\"computers\",\"sustainable agriculture\",\"AI\",\"Customer Support\",\"Network Management\",\"mental health\",\"5G\",\"Data Security\",\"funny stories\",\"sustainable fashion\",\"smart cities\",\"technology\",\"Ecology\",\"health\",\"adventure\",\"App4\",\"Fun\",\"Summary\",\"space exploration\",\"\\\"training\\\"]\",\"artificial intelligence\"]}",
"role": "tool",
"tool_call_id": "tooluse_8ZEYkp2qS_aBpx5cR9XDzA"
}
]
[
{
"name": "Conduit-article-management-getAllTagsForArticles",
"payload_property1": null,
"payload_property2": null,
"raw_response": "{\"tags\":[\"Pet Adventures\",\"BYD\",\"Xpander\",\"dogs\",\"nutrition\",\"Culture\",\"test-oren\",\"test\",\"workshop\",\"xpander.ai\",\"xpanderAI Agent\",\"Lucky\",\"Generative AI\",\"culinary\",\"quantum computing\",\"Rosy\",\"workflow automation\",\"xpanderAI\",\"pet\",\"Issues\",\"example\",\"gig economy\",\"business\",\"??????\",\"electric vehicle\",\"automotive\",\"AWS\",\"ert\",\"Leo\",\"AI Integration\",\"climate change\",\"animals\",\"urban planning\",\"pets\",\"Freddy\",\"fruits\",\"GitHub\",\"augmented reality\",\"conference\",\"Data Pipelines\",\"remote work\",\"history\",\"Testing\",\"animal\",\"Jewel\",\"DriveNets\",\"Nature\",\"Customer Service\",\"Bird\",\"tag1,tag2\",\"Machine Learning\",\"tag1\",\"Collaboration\",\"Lizard\",\"success\",\"apples\",\"adventures\",\"biotechnology\",\"Data Science\",\"gaming\",\"story\",\"computers\",\"sustainable agriculture\",\"AI\",\"Customer Support\",\"Network Management\",\"mental health\",\"5G\",\"Data Security\",\"funny stories\",\"sustainable fashion\",\"smart cities\",\"technology\",\"Ecology\",\"health\",\"adventure\",\"App4\",\"Fun\",\"Summary\",\"space exploration\",\"\\\"training\\\"]\",\"artificial intelligence\"]}",
"response_message": "{\"tags\":[\"Pet Adventures\",\"BYD\",\"Xpander\",\"dogs\",\"nutrition\",\"Culture\",\"test-oren\",\"test\",\"workshop\",\"xpander.ai\",\"xpanderAI Agent\",\"Lucky\",\"Generative AI\",\"culinary\",\"quantum computing\",\"Rosy\",\"workflow automation\",\"xpanderAI\",\"pet\",\"Issues\",\"example\",\"gig economy\",\"business\",\"??????\",\"electric vehicle\",\"automotive\",\"AWS\",\"ert\",\"Leo\",\"AI Integration\",\"climate change\",\"animals\",\"urban planning\",\"pets\",\"Freddy\",\"fruits\",\"GitHub\",\"augmented reality\",\"conference\",\"Data Pipelines\",\"remote work\",\"history\",\"Testing\",\"animal\",\"Jewel\",\"DriveNets\",\"Nature\",\"Customer Service\",\"Bird\",\"tag1,tag2\",\"Machine Learning\",\"tag1\",\"Collaboration\",\"Lizard\",\"success\",\"apples\",\"adventures\",\"biotechnology\",\"Data Science\",\"gaming\",\"story\",\"computers\",\"sustainable agriculture\",\"AI\",\"Customer Support\",\"Network Management\",\"mental health\",\"5G\",\"Data Security\",\"funny stories\",\"sustainable fashion\",\"smart cities\",\"technology\",\"Ecology\",\"health\",\"adventure\",\"App4\",\"Fun\",\"Summary\",\"space exploration\",\"\\\"training\\\"]\",\"artificial intelligence\"]}",
"role": "tool",
"tool_call_id": "tooluse_8ZEYkp2qS_aBpx5cR9XDzA"
}
]
Local tool invocation response (List[ToolResponse])
in case of local tools, invocation should occur manually
[
{
"toolCallId": [
{
"id": "call_U9cbQK8FO8tY97MuqI6mH0dr",
"type": "function",
"function": {
"name": "Conduit-article-management-getAllTagsForArticles",
"arguments": "{}"
}
},
{
"id": "call_jjZgVbbW8FQqhilB5ud54jUJ",
"type": "function",
"function": {
"name": "greet-by-name",
"arguments": "{\"userName\": \"David\"}"
}
}
],
"role": {
"userName": "David"
},
"name": "tool",
"responseMessage": "",
"filteredTool": {},
"payloadRequest": "{\"userName\":\"David\"}"
}
]
[
{
"toolCallId": [
{
"id": "call_U9cbQK8FO8tY97MuqI6mH0dr",
"type": "function",
"function": {
"name": "Conduit-article-management-getAllTagsForArticles",
"arguments": "{}"
}
},
{
"id": "call_jjZgVbbW8FQqhilB5ud54jUJ",
"type": "function",
"function": {
"name": "greet-by-name",
"arguments": "{\"userName\": \"David\"}"
}
}
],
"role": {
"userName": "David"
},
"name": "tool",
"responseMessage": "",
"filteredTool": {},
"payloadRequest": "{\"userName\":\"David\"}"
}
]
[
{
"toolCallId": "tooluse_Ik0QhuxcRAqAsT6P9o5_7w",
"role": "",
"name": "tool",
"responseMessage": "",
"filteredTool": {},
"payloadRequest": "{\"userName\":\"David\"}"
}
]
Manual Tool Invocation
In some scenarios, LLM parsing is not required, and you may prefer to manually invoke functions that exist within the xpander connector.
Request
# API Reference https://docs.xpander.ai/reference/python-sdk#xpander-sdk.XpanderClient.xpanderSingleToolInvoke
payload = {
"body_params":{
"BODY_PARAM": "VALUE"
},
"path_params":{
"PATH_PARAM": "VALUE"
},
"query_params":{
"QUERY_PARAM": "VALUE"
}
}
xpander_client.xpander_single_tool_invoke(tool_id="TOOL_ID_TO_INVOKE",payload=tool_invocation_payload)
# API Reference https://docs.xpander.ai/reference/nodejs-sdk#xpander-sdk.XpanderClient.xpanderSingleToolInvoke
const payload = {
body_params:{
BODY_PARAM: "VALUE"
},
path_params:{
PATH_PARAM: "VALUE"
},
query_params:{
QUERY_PARAM: "VALUE"
}
}
const toolResponse = xpanderClient.xpanderSingleToolInvoke("TOOL_ID_TO_INVOKE", payload);
var payload = new Dictionary<string, object>
{
{ "body_params", new Dictionary<string, string>
{
{ "BODY_PARAM", "VALUE" }
}
},
{ "path_params", new Dictionary<string, string>
{
{ "PATH_PARAM", "VALUE" }
}
},
{ "query_params", new Dictionary<string, string>
{
{ "QUERY_PARAM", "VALUE" }
}
}
};
ToolResponse toolResponse = xpanderClient.XpanderSingleToolInvoke("TOOL_ID_TO_INVOKE", JsonConvert.SerializeObject(payload));
Map<String, Map<String, String>> payload = new HashMap<>();
Map<String, String> bodyParams = new HashMap<>();
bodyParams.put("BODY_PARAM", "VALUE");
Map<String, String> pathParams = new HashMap<>();
pathParams.put("PATH_PARAM", "VALUE");
Map<String, String> queryParams = new HashMap<>();
queryParams.put("QUERY_PARAM", "VALUE");
payload.put("body_params", bodyParams);
payload.put("path_params", pathParams);
payload.put("query_params", queryParams);
ToolResponse[] toolResponses = xpanderClient.xpanderSingleToolInvoke("TOOL_ID_TO_INVOKE", payload);
Response ([ToolResponse])
{
"name": "INVOKED_TOOL_ID",
"payload_property1": null,
"payload_property2": null,
"raw_response": "YOUR_TOOL_RAW_RESPONSE",
"response_message": "YOUR_TOOL_RESPONSE",
"role": "tool",
"tool_call_id": ""
}
[
{
"name": "INVOKED_TOOL_ID",
"payload_property1": null,
"payload_property2": null,
"raw_response": "YOUR_TOOL_RAW_RESPONSE",
"response_message": "YOUR_TOOL_RESPONSE",
"role": "tool",
"tool_call_id": ""
}
]
[
{
"name": "INVOKED_TOOL_ID",
"payload_property1": null,
"payload_property2": null,
"raw_response": "YOUR_TOOL_RAW_RESPONSE",
"response_message": "YOUR_TOOL_RESPONSE",
"role": "tool",
"tool_call_id": ""
}
]