PromptDialog 2.0 coming soon

An Agentic AI Approach to Chatbots

Built on Mica , free on cloud and on-prem.

weather_agent:
type: llm agent
description: Get the current weather in a given location and send email
prompt: |-
You are a Weather Query Agent. You can do following things:
1、Get the current weather in a given location. Location MUST be a city.
When a user asks about the weather, extract the location from their question. Once all the information is collected, call the "get_weather" function
2、sending email:
When a user sending email, extract the recipient、subject and body from their question. Once all the information is collected, call the "send_email" function.
args:
- subject
- recipient
- location
- body
uses:
- get_weather
- send_email
get_weather:
type: function
send_email:
type: function
meta:
type: ensemble agent
description: You can select an agent to response user's question.
contains:
- weather_agent
steps:
- bot: Do you want to know the weather?
fallback:
policy: Lame! We got a glitch, please try again later.
main:
steps:
- call: meta
schedule: priority
import json
from swarm import Agent
def get_weather(location, time="now"):
"""Get the current weather in a given location. Location MUST be a city."""
return json.dumps({"location": location, "temperature": "65", "time": time})
def send_email(recipient, subject, body):
print("Sending email...")
print(f"To: {recipient}")
print(f"Subject: {subject}")
print(f"Body: {body}")
return "Sent!"
weather_agent = Agent(
name="Weather Agent",
instructions="You are a helpful agent.",
functions=[get_weather, send_email],
)
  • All-in-one Studio

    Slots and agents are fully displayed within the same integrated development environment (IDE). Different agents can be flexibly combined and deployed with a single click. The platform supports both Web and Mobile modes for publishing. Simply copy a few lines of code to integrate conversational capabilities into your applications.
  • Agent First Design

    Four types of agents—knowledge base, LLM, flow, and ensemble agents—are predefined for your customization. FAQs and unstructured content stored in Word, PDF, and HTML can be directly indexed for question answering without annotation.
  • Fast Debugging

    It is the first system of its kind to embed the software development process into chatbot design and debugging. Each agent can be tested individually or together to facilitate debugging.

PromptDialog Studio

All-in-one design environment

Layer 1Layer 2