Action-enabled AI

Action-enabled AI

Automatically generated APIs for Agents

Leverage auto-generated graph systems and function calling to seamlessly integrate your APIs and unlock the full potential of your AI applications in minutes.

Don’t work hard on adjusting your APIs for Agentic applications

Let the xpander platform integrate your AI Apps to your APIs in minutes

Connect any AI framework to any SaaS or home-grown API

  • BEDROCK
    CONVERSE

  • AMAZON Q

  • NVIDIA NIM

  • 🦜🔗 LANGCHAIN

Build AI Agents that Perform Real Actions

AI-ready Connectors with Graph

Connect your AI applications to popular SaaS platforms and home-grown systems with xpander’s AI-ready connectors.

AI-ready Connectors with Graph

Connect your AI applications to popular SaaS platforms and home-grown systems with xpander’s AI-ready connectors.

AI-ready Connectors with Graph

Connect your AI applications to popular SaaS platforms and home-grown systems with xpander’s AI-ready connectors.

AI-ready Connectors with Graph

Connect your AI applications to popular SaaS platforms and home-grown systems with xpander’s AI-ready connectors.

AI-ready Connectors with Graph

Connect your AI applications to popular SaaS platforms and home-grown systems with xpander’s AI-ready connectors.

Agent-oriented Application Development

Build and optimize AI agents quickly using xpander’s App workbench and Graph-assisted actions.

Agent-oriented Application Development

Build and optimize AI agents quickly using xpander’s App workbench and Graph-assisted actions.

Agent-oriented Application Development

Build and optimize AI agents quickly using xpander’s App workbench and Graph-assisted actions.

Agent-oriented Application Development

Build and optimize AI agents quickly using xpander’s App workbench and Graph-assisted actions.

Agent-oriented Application Development

Build and optimize AI agents quickly using xpander’s App workbench and Graph-assisted actions.

Use reliable AI Agents as Assistants, Automations, or within your product

Easily apply your own use-case to create a practical agent.

Use reliable AI Agents as Assistants, Automations, or within your product

Easily apply your own use-case to create a practical agent.

Use reliable AI Agents as Assistants, Automations, or within your product

Easily apply your own use-case to create a practical agent.

Use reliable AI Agents as Assistants, Automations, or within your product

Easily apply your own use-case to create a practical agent.

Use reliable AI Agents as Assistants, Automations, or within your product

Easily apply your own use-case to create a practical agent.

What is xpander AI-Ready connector?

What is xpander AI-Ready connector?

xpander connectors are generated by a multi-agent pipeline which ingests data from multiple sources like existing OpenAPI Specs, API documentation, user traffic, and browsing recordings. This pipeline automatically generates an AI-ready connector, which includes a directed graph for optimal API operation dependency fulfillment, as well as OpenAPI Specs with enriched OperationIds, Descriptions, Cross-operation Dependencies, Examples, and Parameters

xpander connectors are generated by a multi-agent pipeline which ingests data from multiple sources like existing OpenAPI Specs, API documentation, user traffic, and browsing recordings. This pipeline automatically generates an AI-ready connector, which includes a directed graph for optimal API operation dependency fulfillment, as well as OpenAPI Specs with enriched OperationIds, Descriptions, Cross-operation Dependencies, Examples, and Parameters

Stop wasting time on writing function calls

Stop wasting time on writing function calls

Agentic hello-world with xpander

agentic_prompt = f"A user named '{username}' has just signed in. 
Would you like to provide a personalized greeting for the user?"

messages = [
		{"role": "user", "content": agentic_prompt}
	]

# Retrieve tools for the OpenAI plugin from the client
tools = xpander_client.tools()

# Create a chat completion using the OpenAI client
llm_response = openai_client.chat.completions.create(
	model="gpt-4o",
	messages=messages,
	tools=tools,
	tool_choice="required"
)

# LLM Decided to invoke #Hello world function from the API
xpander_client.xpander_tool_call(llm_response)
agentic_prompt = f"A user named '{username}' has just signed in. 
Would you like to provide a personalized greeting for the user?"

messages = [
		{"role": "user", "content": agentic_prompt}
	]

# Retrieve tools for the OpenAI plugin from the client
tools = xpander_client.tools()

# Create a chat completion using the OpenAI client
llm_response = openai_client.chat.completions.create(
	model="gpt-4o",
	messages=messages,
	tools=tools,
	tool_choice="auto"
)

# LLM Decided to invoke #Hello world function from the API
xpander_client.xpander_tool_call(llm_response)

Agentic hello-world without xpander

def get_hello_world_message(username):
    response = requests.post('https://example.com/api/hello', json=username)
    return response

agentic_prompt = f"A user named '{username}' has just signed in. 
Would you like to provide a personalized greeting for the user?"

messages = [
		{"role": "user", "content": agentic_prompt}
	]

# manualy specify the tools
tools = [{
          "type": "function",
          "function": {
              "name": "get_hello_world_message",
              "parameters": {
                  "type": "object",
                  "properties": {
                      "username": {"type": "string"}
                  },
                  "required": ["username"]
              }
          }
		  ## Add more functions here
      }]

# Create a chat completion using the OpenAI client
llm_response = openai_client.chat.completions.create(
	model="gpt-4o",
	messages=messages,
	tools=tools,
	tool_choice="required"
)

# Executing the response 
for choice in llm_response.choices:
	if choice.finish_reason == "tool_calls":
		for tool_call in choice.message.tool_calls:
			if tool_call.function.name == "get_hello_world_message":
				get_hello_world_message(**params)
			## If function 2
			## If function 3
			## If function ..
def get_hello_world_message(username):
    response = requests.post('https://example.com/api/hello', json=username)
    return response

agentic_prompt = f"A user named '{username}' has just signed in. 
Would you like to provide a personalized greeting for the user?"

messages = [
		{"role": "user", "content": agentic_prompt}
	]

# manualy specify the tools
tools = [{
          "type": "function",
          "function": {
              "name": "get_hello_world_message",
              "parameters": {
                  "type": "object",
                  "properties": {
                      "username": {"type": "string"}
                  },
                  "required": ["username"]
              }
          }
		  ## Add more functions here
      }]

# Create a chat completion using the OpenAI client
llm_response = openai_client.chat.completions.create(
	model="gpt-4o",
	messages=messages,
	tools=tools,
	tool_choice="auto"
)

# Executing the response 
for choice in llm_response.choices:
	if choice.finish_reason == "tool_calls":
		for tool_call in choice.message.tool_calls:
			if tool_call.function.name == "get_hello_world_message":
				get_hello_world_message(**params)
			## If function 2
			## If function 3
			## If function ..