From e7391ff6e10c0d6d30bbfb034bd62f5ec50c9541 Mon Sep 17 00:00:00 2001 From: ascender1729 Date: Thu, 1 May 2025 18:28:53 +0530 Subject: [PATCH] fix: update vision.md with improved agent examples and documentation --- docs/swarms/concept/vision.md | 67 ++++++++++++----------------------- 1 file changed, 22 insertions(+), 45 deletions(-) diff --git a/docs/swarms/concept/vision.md b/docs/swarms/concept/vision.md index 80185bb3..678b495d 100644 --- a/docs/swarms/concept/vision.md +++ b/docs/swarms/concept/vision.md @@ -15,30 +15,14 @@ Swarms aims to be the definitive and most reliable multi-agent LLM framework, of This example demonstrates a simple financial agent setup that responds to financial questions, such as establishing a ROTH IRA, using OpenAI's GPT-based model. ```python -import os -from swarms import Agent -from swarm_models import OpenAIChat -from swarms.prompts.finance_agent_sys_prompt import ( - FINANCIAL_AGENT_SYS_PROMPT, -) -from dotenv import load_dotenv - -# Load environment variables -load_dotenv() - -# Get OpenAI API key from environment -api_key = os.getenv("OPENAI_API_KEY") +from swarms.structs.agent import Agent +from swarms.prompts.finance_agent_sys_prompt import FINANCIAL_AGENT_SYS_PROMPT -# Initialize OpenAIChat model with desired parameters -model = OpenAIChat( - openai_api_key=api_key, model_name="gpt-4o-mini", temperature=0.1 -) - -# Initialize the Financial Analysis Agent +# Initialize the Financial Analysis Agent with GPT-4o-mini model agent = Agent( agent_name="Financial-Analysis-Agent", system_prompt=FINANCIAL_AGENT_SYS_PROMPT, - llm=model, + model_name="gpt-4o-mini", max_loops=1, autosave=True, dashboard=False, @@ -56,7 +40,7 @@ out = agent.run( "How can I establish a ROTH IRA to buy stocks and get a tax break? What are the criteria?" ) -# Output the agent's result +# Output the result print(out) ``` @@ -64,66 +48,60 @@ print(out) The following example showcases how to use the `AgentRearrange` class to manage a multi-agent system. It sets up a director agent to orchestrate two workers—one to generate a transcript and another to summarize it. ```python -from swarms import Agent, AgentRearrange -from swarm_models import Anthropic +from swarms.structs.agent import Agent +from swarms.structs.rearrange import AgentRearrange -# Initialize the Director agent +# Initialize the Director agent using Anthropic model via model_name director = Agent( agent_name="Director", - system_prompt="Directs the tasks for the workers", - llm=Anthropic(), + system_prompt="You are a Director agent. Your role is to coordinate and direct tasks for worker agents. Break down complex tasks into clear, actionable steps.", + model_name="claude-3-sonnet-20240229", max_loops=1, dashboard=False, - streaming_on=True, + streaming_on=False, verbose=True, stopping_token="", state_save_file_type="json", saved_state_path="director.json", ) -# Initialize Worker 1 agent (transcript generation) +# Worker 1: transcript generation worker1 = Agent( agent_name="Worker1", - system_prompt="Generates a transcript for a YouTube video on what swarms are", - llm=Anthropic(), + system_prompt="You are a content creator agent. Your role is to generate detailed, engaging transcripts for YouTube videos about technical topics. Focus on clarity and educational value.", + model_name="claude-3-sonnet-20240229", max_loops=1, dashboard=False, - streaming_on=True, + streaming_on=False, verbose=True, stopping_token="", state_save_file_type="json", saved_state_path="worker1.json", ) -# Initialize Worker 2 agent (summarizes transcript) +# Worker 2: summarization worker2 = Agent( agent_name="Worker2", - system_prompt="Summarizes the transcript generated by Worker1", - llm=Anthropic(), + system_prompt="You are a summarization agent. Your role is to create concise, clear summaries of technical content while maintaining key information and insights.", + model_name="claude-3-sonnet-20240229", max_loops=1, dashboard=False, - streaming_on=True, + streaming_on=False, verbose=True, stopping_token="", state_save_file_type="json", saved_state_path="worker2.json", ) -# Create a list of agents +# Orchestrate the agents in sequence agents = [director, worker1, worker2] - -# Define the workflow pattern (sequential flow) flow = "Director -> Worker1 -> Worker2" - -# Using AgentRearrange to orchestrate the agents agent_system = AgentRearrange(agents=agents, flow=flow) -# Running the system with a sample task +# Run the workflow output = agent_system.run( "Create a format to express and communicate swarms of LLMs in a structured manner for YouTube" ) - -# Output the result print(output) ``` @@ -168,5 +146,4 @@ With support for extreme third-party integration, Swarms makes it easy for devel Swarms abstracts the complexity of managing multiple agents with orchestration tools like `AgentRearrange`. Developers can define workflows that execute tasks concurrently or sequentially, depending on the problem at hand. This makes it easy to build and maintain large-scale automation systems. ### Conclusion: -Swarms is not just another multi-agent framework; it's built specifically for developers who need powerful tools to automate complex, large-scale business operations. With flexible architecture, deep integration capabilities, and developer-friendly APIs, Swarms is the ultimate solution for businesses looking to streamline operations and future-proof their workflows. - +Swarms is not just another multi-agent framework; it's built specifically for developers who need powerful tools to automate complex, large-scale business operations. With flexible architecture, deep integration capabilities, and developer-friendly APIs, Swarms is the ultimate solution for businesses looking to streamline operations and future-proof their workflows. \ No newline at end of file