Skip to content

Dynamically Accessing AI Assistant Metadata

Instead of manually configuring AI Assistant URLs, you can use the Practicus AI SDK to dynamically fetch available AI Assistants associated with your user account.

import json
import practicuscore as prt

region = prt.current_region()
ai_assistants = region.get_ai_assistants()

print("Available AI Assistants:")
for assistant in ai_assistants:
    print(json.dumps(assistant, indent=4))

Example: Selecting the First AI Assistant

selected_ai_assistant = ai_assistants[0]

from practicuscore.gen_ai import AIAssistantHelper

assistant_url = AIAssistantHelper.get_api_endpoint(selected_ai_assistant, region)
assistant_token = None  # Get a new token, or reuse existing if not expired.
assistant_token = AIAssistantHelper.get_api_token(
    assistant_config=selected_ai_assistant, region=region, token=assistant_token
)

print("Use the following URL for API requests:")
print(assistant_url)

print("API Interface Type (Auto, OpenAI, or Langchain):")
print(selected_ai_assistant["api_interface"])

print("Bearer Token for API authentication:")
print(assistant_token)

Previous: Lang Chain LLM Model | Next: Mobile Banking > Mobile-Banking