crewAI
crewAI copied to clipboard
crewAIInc
Add Google Gemini API Support
Since Google Gemini Pro API is currently free when using up to 60 API calls per minute. This would be an incredibly helpful integration to add support for Gemini API into the CrewAI code. This will perform better than GPT 3.5 without any API fees. However, CrewAI must be able to simply allow users to define an API call/Functions Limit, so they can set it to 59 API processes per minute to avoid going over the free tier use per minute limits.
import os from crewai import Agent, Task, Crew, Process from langchain_google_genai import ChatGoogleGenerativeAI
llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.1,google_api_key="GEMINI-API-KEY")
Define your agents with roles and goals
researcher = Agent( role='Senior Research Analyst', goal='Uncover cutting-edge developments in AI and data science', backstory="""You work at a leading tech think tank. Your expertise lies in identifying emerging trends. You have a knack for dissecting complex data and presenting actionable insights.""", verbose=True, llm = llm, allow_delegation=False, tools=[], )
writer = Agent( role='Tech Content Strategist', goal='Craft compelling content on tech advancements', backstory="""You are a renowned Content Strategist, known for your insightful and engaging articles. You transform complex concepts into compelling narratives.""", verbose=True, allow_delegation=True, llm = llm, tools=[], )
task1 = Task( description="""Conduct a comprehensive analysis of the latest advancements in AI in 2024. Identify key trends, breakthrough technologies, and potential industry impacts. Your final answer MUST be a full analysis report""", agent=researcher )
task2 = Task( description="""Using the insights provided, develop an engaging blog post that highlights the most significant AI advancements. Your post should be informative yet accessible, catering to a tech-savvy audience. Make it sound cool, avoid complex words so it doesn't sound like AI. Your final answer MUST be the full blog post of at least 4 paragraphs.""", agent=writer )
Instantiate your crew with a sequential process
crew = Crew( agents=[researcher, writer], tasks=[task1, task2], verbose=2, )
Get your crew to work!
result = crew.kickoff()
print("######################") print(result)
##Thank you for the update
@janda-datascience Thank you so much! Really excited to get CrewAI working with Google Gemini. I ran into one issue with your code, it appers /crewai/agent.py is still looking for openAI and its key for its functions and won't run without an OpenAI key with the error: "Did not find openai_api_key, please add an environment variable OPENAI_API_KEY which contains it".
I assume settings need to be changed in /crewai/agent.py to point to the Gemini Pro API instead of OpenAI for all agent functions, so it doesn't bill the OpenAI API? It looks like this is the section, could you please help provide the proper updates to agent.py . In contrast, If I just add the OpenAI API, it start running agent tasks and billing me on openai api. Really appreciate your help.
agent.py section to modify? (Lines 57-60 in agent.py)
llm: Optional[Any] = Field( default_factory=lambda: ChatOpenAI( temperature=0.7, model_name="gpt-4",
Error: agent.py calling for OpenAI still:
`Traceback (most recent call last):
File "/home/aivirtual/apps/CrewAIGoogleGemini/CrewAI-TEST-2.py", line 41, in
File "/home/aivirtual/miniconda3/envs/crewai/lib/python3.10/site-packages/langchain_core/load/serializable.py", line 107, in init
super().init(**kwargs)
File "/home/aivirtual/miniconda3/envs/crewai/lib/python3.10/site-packages/pydantic/v1/main.py", line 341, in init
raise validation_error
pydantic.v1.error_wrappers.ValidationError: 1 validation error for ChatOpenAI
root
Did not find openai_api_key, please add an environment variable OPENAI_API_KEY which contains it, or pass openai_api_key as a named parameter. (type=value_error)`
@mindwellsolutions by default the agent uses gpt-4, if you want to use your gemini model you can provide the gemini llm inside the agent definition:
eg:
from langchain_google_genai import ChatGoogleGenerativeAI
from crewai import Agent, Task, Crew, Process
#this is gemini llm
llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.1)
#using gemini llm inside agent
SQLdev = Agent(
role='SQL Developer ',
goal='Give best Possible solution along with code',
backstory="You are experienced sql developer who always provide optimized query",
verbose=True,
allow_delegation=False,
llm = llm, #using google gemini
tools=[
BrowserTools.scrape_and_summarize_website,
SearchTools.search_internet,
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
@punitchauhan771 Thank you so much, that worked! It successfully runs CrewAI completely free with Gemini Pro API! Going to spend some time testing it out. Really appreciate the help here. I had to add one small update to the solution punitchauhan provided to get it working, thought I'd share it with everyone. You need to define the google_api_key directly within the llm = line of code.
(Working Update to llm = code for gemini api key)
llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.1,google_api_key="<Enter Google Gemini API KEY>")
Last Question: Google Gemini API (Free) has a limit of 60 API uses per minute. Is there a way to throttle the agent activity, so it doesn't perform more than 59 API requests per minute, so we can avoid going over free limits when dispatching large agent swarms? Really appreciate everything, ty.
Hi @mindwellsolutions, I don't think there is a way to throttle the agent activity, However if you want to know how many request did your agent made you can use langchain module example code:
import os
from crewai import Agent, Task, Crew, Process
from langchain.callbacks import get_openai_callback
# Define your agents with roles and goals
CopyWriter = Agent(
role='Copy Writer',
goal='To write the best article',
backstory="You're an experienced Copy writer who writes Technical articles",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
BrowserTools.scrape_and_summarize_website,
SearchTools.search_internet,
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
SEO_Researcher = Agent(
role='SEO analyst',
goal='To Give the best seo based analyst tags',
backstory="You're an experienced seo analyst who Give the best seo tags",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
BrowserTools.scrape_and_summarize_website,
SearchTools.search_internet,
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
# Define your agents with roles and goals
researcher = Agent(
role='Researcher',
goal='You provide results on the basis of Facts and only Facts along with supported doc related URLs ,You go to the root cause and give the best possible outcomes',
backstory="You're an ai researcher who researches on the field of AI and have won multiple awards",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
BrowserTools.scrape_and_summarize_website,
SearchTools.search_internet,
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
# Create tasks for your agents
task1 = Task(description='reseach on crew ai, reference url = https://github.com/joaomdmoura/crewAI', agent=researcher)
task2 = Task(description=f'Create an article on langchain agents and tools and also give an example and also write a detail summary on the basis of {researcher} response', agent = CopyWriter)
task3 = Task(description=f'give me best tags for the article written by {CopyWriter} Agent', agent=SEO_Researcher)
# Instantiate your crew with a sequential process
crew = Crew(
agents=[researcher,CopyWriter,SEO_Researcher],
tasks=[task1,task2,task3],
verbose=2, # Crew verbose more will let you know what tasks are being worked on, you can set it to 1 or 2 to different logging levels
process=Process.sequential # Sequential process will have tasks executed one after the other and the outcome of the previous one is passed as extra content into this next.
)
# result = crew.kickoff()
# Get your crew to work!
with get_openai_callback() as cb:
result = crew.kickoff()
print(result)
print(cb)
at the end of the agent activity it will give you some response like :
Tokens Used: 0
Prompt Tokens: 0
Completion Tokens: 0
Successful Requests: 11
Total Cost (USD): $0.0
hope this helps.
@punitchauhan771 Thank you and everyone for the help, such an amazing community. This would help a lot, but I'm using Google Gemini API rather than OpenAI with from langchain_google_genai import ChatGoogleGenerativeAI.
Is there a callback for ChatGoogleGenerativeAI similar to the OpenAI callback you posited "from langchain.callbacks import get_openai_callback". Really appreciate the help.
Hi @mindwellsolutions the solution I provided works for Gemini as well,though it doesn't count tokens 🙂.
@punitchauhan771 Thanks again. Unfortunately using your solution after 2 agents complete all their tasks successfully with Gemini API its still showing "Successful Request: 0". As you mentioned simply being able to monitor the amount of Successful Requests is all Gemini Pro (Free Tier) needs since there are no costs to track associated with token usage :)
The Output I get after Gemini clearly performs multiple tasks:
Tokens Used: 0 Prompt Tokens: 0 Completion Tokens: 0 Successful Requests: 0 Total Cost (USD): $0.0
@mindwellsolutions if possible can you provide me the code snippet? Because I just ran the code and this is the response I got while I used Gemini.
Tokens Used: 0
Prompt Tokens: 0
Completion Tokens: 0
Successful Requests: 11
Total Cost (USD): $0.0
@punitchauhan771 Actually. I closed the terminal window in VS Code and ran it again and it worked perfect. Really appreciate all your help, you've gotten everything running perfectly for me. Ty
The performance of the Gemini Pro model using tools is not good. There are big problems with the logic and results of using the tools, and the final results are often hallucinatory.
@edisonzf2020 Thanks for your comment. I got time to test out gemini api in crewai further over the weekend and as you mentioned it seems to be having issues using tools like duckduckgosearch. I tested openai API and Ollama (zephyr) and those both worked perfectly, while it looks like Gemini is pulling responses from its internal knowledgebase rather than relaying the data from DuckDuckGo.
Are there any possible fixes moving forward to get Gemini to play well with CrewAI? The scalability of using a free API for personal use will provide incredible power of scalability without having to run local models on local GPU resources.
Hey folks, catching up to this issue! great comments, so glad you were able to get Gemini working, we are adding new docs that will have instructions for all the major models so stay tuned for that.
I'll do some testing with Gemini models specifically to see how we could make that better!
@joaomdmoura Thank you. This is incredibly appreciated. We realized that Gemini API has significant problems currently using CrewAI's tools like DuckDuckGoSearch. It appears none of the data from the tools make it back to Gemini, and Gemini always generates answers from its internal knowledge base rather than the research done by the tool. If there is a way to fix this so Gemini works like other LLMs that would be amazing.
I've been using Zephyr 7b as a local model and that runs well, but the power of decent size scalability of swarms that would require significant power to run locally, can be run from any device using Gemini's free API - which is going to be extremely valuable. Especially if each Agent is assigned it's own Free Gemini API key in concert with each other, the ability to built large API swarms for free will be significant.
Thanks for all that you do. Really loving the functionality of CrewAI.
@edisonzf2020 Thanks for your comment. I got time to test out gemini api in crewai further over the weekend and as you mentioned it seems to be having issues using tools like duckduckgosearch. I tested openai API and Ollama (zephyr) and those both worked perfectly, while it looks like Gemini is pulling responses from its internal knowledgebase rather than relaying the data from DuckDuckGo.
Are there any possible fixes moving forward to get Gemini to play well with CrewAI? The scalability of using a free API for personal use will provide incredible power of scalability without having to run local models on local GPU resources.
Hi @mindwellsolutions, I’m uncertain if it’s sourcing answers from its internal knowledge base . For reference, I’ve attempted the same inquiry using the Gemini LLM, Agent, and CrewAI Agent. If it’s utilizing its internal knowledge base, then it shouldn’t need tools like search, right? However, when I tried using the CrewAI Agent without any search tool, it provided me with this response. code:
researcher = Agent(
role='Researcher',
goal='You provide results on the basis of Facts and only Facts along with supported doc related urls,You go to the root cause and give the best possible outcomes',
backstory="You're an ai researcher who researches on the field of AI and have won multiple awards",
verbose=True,
allow_delegation=False,
llm = llm)
task1 = Task(description='research on crewai in context of llm agents', agent=researcher)
crew.kickoff()
response:
> Entering new CrewAgentExecutor chain...
Thought: Do I need to use a tool? Yes
Action: Google Search
Action Input: crewai in context of llm agentsGoogle Search is not a valid tool, try one of [].Do I need to use a tool? No
Final Answer: I am sorry, I do not have access to the internet to perform a Google search on crewai in the context of LLM agents.
> Finished chain.
[DEBUG]: [Researcher] Task output: I am sorry, I do not have access to the internet to perform a Google search on crewai in the context of LLM agents.
I am sorry, I do not have access to the internet to perform a Google search on crewai in the context of LLM agents.
when i tried google search the response was:
> Entering new CrewAgentExecutor chain...
Thought: Do I need to use a tool? Yes
Action: Search the internet
Action Input: crewai in context of llm agents{'searchParameters': {'q': 'crewai in context of llm agents', 'type': 'search', 'engine': 'google'}, 'organic': [{'title': 'CrewAI: A Team of AI Agents that Work Together for You - Medium', 'link': 'https://medium.com/@mayaakim/crewai-a-team-of-ai-agents-that-work-together-for-you-4cc9d24e0857', 'snippet': 'In this really cool intro to LLMs, Andrej, one of the top engineers at OpenAI references a book by Daniel Kahneman “Thinking Fast and Slow”.', 'date': 'Jan 5, 2024', 'position': 1}, {'title': 'CrewAI Unleashed: Future of AI Agent Teams - LangChain Blog', 'link': 'https://blog.langchain.dev/crewai-unleashed-future-of-ai-agent-teams/', 'snippet': 'AI Agents Crews are game-changing. AI agents are emerging as game-changers, quickly becoming partners in problem-solving, creativity, and ...', 'date': 'Dec 21, 2023', 'position': 2}, {'title': 'CrewAI - Your Own Team of Autonomous Agents - YouTube', 'link': 'https://www.youtube.com/watch?v=B8s-FyN4UeE', 'snippet': 'CrewAI is an autonomous Agents framework that enable you to build a crew of agents to ...', 'date': 'Jan 13, 2024', 'attributes': {'Duration': '10:32', 'Posted': 'Jan 13, 2024'}, 'imageUrl': 'https://i.ytimg.com/vi/B8s-FyN4UeE/default.jpg?sqp=-oaymwEECHgQQw&rs=AMzJL3k9fjeL4ynDy_DtrDvC3Edp-SJ5dQ', 'position': 3}, {'title': 'Crew AI | Unleashing the Potential of Multi-Agent Interactions', 'link': 'https://medium.com/@rajib76.gcp/crew-ai-unleashing-the-potential-of-multi-agent-interactions-d57ce8b00fbc', 'snippet': 'This method promotes a modular and complementary amalgamation of the extensive skills provided by LLMs.', 'date': '8 days ago', 'position': 4}, {'title': 'CrewAI: Framework For Creating Autonomous AI Agents - YouTube', 'link': 'https://www.youtube.com/watch?v=Kbq9m-x7gYU', 'snippet': 'In this in-depth tutorial, we dive into the revolutionary CrewAI Framework, focusing on the game ...', 'date': 'Jan 10, 2024', 'attributes': {'Duration': '11:57', 'Posted': 'Jan 10, 2024'}, 'imageUrl': 'https://i.ytimg.com/vi/Kbq9m-x7gYU/default.jpg?sqp=-oaymwEECHgQQw&rs=AMzJL3lQf_a4mE2Yf14VahaxKqkL8MMbDg', 'position': 5}, {'title': 'CrewAi + Solor/Hermes + Langchain + Ollama = Super Ai Agent', 'link': 'https://pub.towardsai.net/crewai-solor-hermes-langchain-ollama-super-ai-agent-0ee348404428', 'snippet': '... explanations; LLM: This stands for “large language model” and in this case, ollama_openhermes is passed as the model for the agent to use.', 'date': '8 days ago', 'position': 6}, {'title': 'Company Spotlight: CrewAI - Replit — Blog', 'link': 'https://blog.replit.com/crew-ai', 'snippet': 'AI agents are here to stay Large Language Models (LLMs) are everywhere, doing various jobs, from chatting to parsing documents.', 'date': 'Dec 21, 2023', 'position': 7}, {'title': 'CrewAI : How To Build AI Agent Teams', 'link': 'https://gptpluginz.com/crewai-how-to-build-ai-agent-teams/', 'snippet': 'Orchestrating role-playing, autonomous AI agents for complex tasks. Development of LLM applications using conversable, customizable agents.', 'date': 'Jan 6, 2024', 'position': 8}, {'title': 'CrewAI: AI-Powered Blogging Agents using LM Studio, Ollama ...', 'link': 'https://www.youtube.com/watch?v=fnchsJd9pfE', 'snippet': "Welcome to an exciting journey into the world of AI-powered blogging! In today's video, I take ...", 'date': '6 days ago', 'attributes': {'Duration': '8:14', 'Posted': '6 days ago'}, 'imageUrl': 'https://i.ytimg.com/vi/fnchsJd9pfE/default.jpg?sqp=-oaymwEECHgQQw&rs=AMzJL3lTSxNEYaT9UykAOakyZkYNY50VUg', 'position': 9}, {'title': 'CrewAI agent framework with local models : r/LocalLLaMA - Reddit', 'link': 'https://www.reddit.com/r/LocalLLaMA/comments/18v527r/crewai_agent_framework_with_local_models/', 'snippet': 'Basically, all the agentic software I want to develop needs tools. I tried connecting to Ollama via LiteLLM (as an OpenAI proxy) which also ...', 'date': 'Dec 31, 2023', 'position': 10}], 'peopleAlsoAsk': [{'question': 'What are different types of agents in AI?', 'snippet': "AGENTS IN ARTIFICIAL INTELLIGENCE CAN BE CATEGORIZED INTO DIFFERENT TYPES BASED ON HOW AGENT'S ACTIONS AFFECT THEIR PERCEIVED INTELLIGENCE AND CAPABILITIES, SUCH AS:\nSimple reflex agents.\nModel-based agents.\nGoal-based agents.\nUtility-based agents.\nLearning agents.\nHierarchical agents.", 'title': '6 Types of AI Agents: Exploring the Future of Intelligent Machines', 'link': 'https://www.simform.com/blog/types-of-ai-agents/'}, {'question': 'What is a model-based reflex agent in AI?', 'snippet': "A model-based reflex agent is one that uses internal memory and a percept history to create a model of the environment in which it's operating and make decisions based on that model. The term percept means something that has been observed or detected by the agent.", 'title': 'Agent-Based Modeling | Process & Examples - Video & Lesson Transcript', 'link': 'https://study.com/academy/lesson/model-based-agents-definition-interactions-examples.html'}, {'question': 'What is a goal-based agent in AI?', 'snippet': 'A goal-based agent is an AI system designed to achieve a specific goal. The goal can be anything from navigating a maze to playing a game. Given a plan, a goal-based agent attempts to choose the best strategy to achieve it based on the environment.', 'title': 'Difference Between Goal-based and Utility-based Agents - Baeldung', 'link': 'https://www.baeldung.com/cs/goal-based-vs-utility-based-agents'}, {'question': 'What is learning agent in artificial intelligence?', 'snippet': 'A learning agent is a tool in AI that is capable of learning from its experiences. It starts with some basic knowledge and is then able to act and adapt autonomously, through learning, to improve its own performance.', 'title': 'Learning Agents: Definition, Components & Examples - Study.com', 'link': 'https://study.com/academy/lesson/learning-agents-definition-components-examples.html'}], 'relatedSearches': [{'query': 'Crewai in context of llm agents examples'}]}
Title: CrewAI: A Team of AI Agents that Work Together for You - Medium
Link: https://medium.com/@mayaakim/crewai-a-team-of-ai-agents-that-work-together-for-you-4cc9d24e0857
Snippet: In this really cool intro to LLMs, Andrej, one of the top engineers at OpenAI references a book by Daniel Kahneman “Thinking Fast and Slow”.
-----------------
Title: CrewAI Unleashed: Future of AI Agent Teams - LangChain Blog
Link: https://blog.langchain.dev/crewai-unleashed-future-of-ai-agent-teams/
Snippet: AI Agents Crews are game-changing. AI agents are emerging as game-changers, quickly becoming partners in problem-solving, creativity, and ...
-----------------
Title: CrewAI - Your Own Team of Autonomous Agents - YouTube
Link: https://www.youtube.com/watch?v=B8s-FyN4UeE
Snippet: CrewAI is an autonomous Agents framework that enable you to build a crew of agents to ...
-----------------
Title: Crew AI | Unleashing the Potential of Multi-Agent Interactions
Link: https://medium.com/@rajib76.gcp/crew-ai-unleashing-the-potential-of-multi-agent-interactions-d57ce8b00fbc
Snippet: This method promotes a modular and complementary amalgamation of the extensive skills provided by LLMs.
-----------------Do I need to use a tool? No
Final Answer: CrewAI is a platform that allows users to create and manage teams of AI agents. These agents can be used to automate tasks, generate content, and provide customer service. CrewAI is powered by large language models (LLMs), which are a type of AI that can understand and generate human language. LLMs are trained on massive datasets of text and code, which allows them to learn to perform a wide variety of tasks.
CrewAI's agents are designed to work together as a team, which allows them to solve problems that would be difficult for a single agent to solve. For example, a team of agents could be used to develop a marketing campaign, write a blog post, or create a customer service chatbot.
CrewAI is a powerful tool that can be used to automate tasks, generate content, and provide customer service. It is a valuable resource for businesses and individuals who want to use AI to improve their productivity and efficiency.
> Finished chain.
[DEBUG]: [Researcher] Task output: CrewAI is a platform that allows users to create and manage teams of AI agents. These agents can be used to automate tasks, generate content, and provide customer service. CrewAI is powered by large language models (LLMs), which are a type of AI that can understand and generate human language. LLMs are trained on massive datasets of text and code, which allows them to learn to perform a wide variety of tasks.
CrewAI's agents are designed to work together as a team, which allows them to solve problems that would be difficult for a single agent to solve. For example, a team of agents could be used to develop a marketing campaign, write a blog post, or create a customer service chatbot.
CrewAI is a powerful tool that can be used to automate tasks, generate content, and provide customer service. It is a valuable resource for businesses and individuals who want to use AI to improve their productivity and efficiency.
CrewAI is a platform that allows users to create and manage teams of AI agents. These agents can be used to automate tasks, generate content, and provide customer service. CrewAI is powered by large language models (LLMs), which are a type of AI that can understand and generate human language. LLMs are trained on massive datasets of text and code, which allows them to learn to perform a wide variety of tasks.\n\nCrewAI's agents are designed to work together as a team, which allows them to solve problems that would be difficult for a single agent to solve. For example, a team of agents could be used to develop a marketing campaign, write a blog post, or create a customer service chatbot.\n\nCrewAI is a powerful tool that can be used to automate tasks, generate content, and provide customer service. It is a valuable resource for businesses and individuals who want to use AI to improve their productivity and efficiency.
and when I tried duckduckgoSearch the response was:
[DEBUG]: Working Agent: Researcher
[INFO]: Starting Task: research on crewai in context of llm agents
> Entering new CrewAgentExecutor chain...
Thought: Do I need to use a tool? Yes
Action: Scrape website content
Action Input: https://crew.ai/The content provided is an advertisement for the sale of the domain name "crew.ai" through Dan.com. It highlights the benefits of using their platform, including buyer protection, fast and easy transfers, and hassle-free payments. The content does not contain any relevant information that can be summarized.Do I need to use a tool? Yes
Action: duckduckgo_search
Action Input: crewai llm agentsCrewAI is a Framework that will make easy for us to get Local AI Agents interacting between them. Using Crew AI Pre-Requisites - Get Docker! 👇 Follow the steps below to get CrewAI in a Docker Container to have all the dependencies contained. With this approach, we will get our Free AI Agents interacting between them locally. CrewAI's vision is clear, to allow engineers to harness the collective power of AI agents, moving beyond traditional automation, by bringing Agents together and allowing for streamlined decision-making, enhanced creativity, and solving complex challenges. Simplicity through Modular Design krazzmann CrewAI agent framework with local models Resources This is great news for everyone who wants to develop agentic software. After a lot of failure and disappointments with running Autogen with local models, I tried the rising star of agent frameworks, CrewAI. CrewAI is a library specifically designed to build and orchestrate groups of AI agents. It's made to be straightforward and modular, so integrating it into your projects is a breeze. Think of CrewAI like a set of building blocks - each piece is unique, but they're all designed to fit together smoothly. CrewAI is a new multi-agent framework built on top of LangChain to simplify LLM-based Agent development. In this blog post, I'll guide you through the essentials of using CrewAI to streamline complex workflows with Python. From setting up AI agents to managing tasks, you'll learn how CrewAI's versatility can enhance efficiency in various domains.Do I need to use a tool? No
Final Answer: CrewAI is a framework that simplifies the development and orchestration of groups of AI agents. It is designed to be straightforward and modular, making it easy to integrate into existing projects. CrewAI is built on top of LangChain, a platform for building and deploying language models. It provides a set of building blocks that can be used to create complex workflows with Python. CrewAI is particularly useful for tasks that require collaboration between multiple AI agents, such as decision-making, creativity, and problem-solving.
Here are some additional details about CrewAI:
* It is open-source and available on GitHub.
* It is compatible with a variety of AI models, including GPT-3, T5, and BERT.
* It provides a range of features for managing AI agents, including task assignment, communication, and monitoring.
* It has been used to develop a variety of applications, including chatbots, customer service agents, and game AI.
Overall, CrewAI is a powerful tool for developing and deploying AI agents. It is easy to use and can be used to create a wide range of applications.
> Finished chain.
[DEBUG]: [Researcher] Task output: CrewAI is a framework that simplifies the development and orchestration of groups of AI agents. It is designed to be straightforward and modular, making it easy to integrate into existing projects. CrewAI is built on top of LangChain, a platform for building and deploying language models. It provides a set of building blocks that can be used to create complex workflows with Python. CrewAI is particularly useful for tasks that require collaboration between multiple AI agents, such as decision-making, creativity, and problem-solving.
Here are some additional details about CrewAI:
* It is open-source and available on GitHub.
* It is compatible with a variety of AI models, including GPT-3, T5, and BERT.
* It provides a range of features for managing AI agents, including task assignment, communication, and monitoring.
* It has been used to develop a variety of applications, including chatbots, customer service agents, and game AI.
Overall, CrewAI is a powerful tool for developing and deploying AI agents. It is easy to use and can be used to create a wide range of applications.
CrewAI is a framework that simplifies the development and orchestration of groups of AI agents. It is designed to be straightforward and modular, making it easy to integrate into existing projects. CrewAI is built on top of LangChain, a platform for building and deploying language models. It provides a set of building blocks that can be used to create complex workflows with Python. CrewAI is particularly useful for tasks that require collaboration between multiple AI agents, such as decision-making, creativity, and problem-solving.\n\nHere are some additional details about CrewAI:\n\n* It is open-source and available on GitHub.\n* It is compatible with a variety of AI models, including GPT-3, T5, and BERT.\n* It provides a range of features for managing AI agents, including task assignment, communication, and monitoring.\n* It has been used to develop a variety of applications, including chatbots, customer service agents, and game AI.\n\nOverall, CrewAI is a powerful tool for developing
also when i tried the same using langchain agent, the response were same, but when i just used the basic llm, the response was:
AIMessage(content="CreAI is a powerful AI-powered writing assistant that helps you create high-quality content quickly and easily. It uses advanced natural language processing (NLP) and machine learning algorithms to understand your writing style and generate text that is both informative and engaging.\n\nWith CreAI, you can:\n\n* **Generate unique and original content:** CreAI can help you create unique and original content that is free of plagiarism. It uses a variety of sources to gather information and then generates text that is both accurate and interesting.\n* **Improve your writing style:** CreAI can help you improve your writing style by identifying common errors and suggesting improvements. It can also help you develop a more consistent and professional writing style.\n* **Save time:** CreAI can help you save time by generating content quickly and easily. This can free up your time to focus on other tasks, such as marketing and promotion.\n\nCreAI is a valuable tool for anyone who wants to create high-quality content quickly and easily. It is especially useful for businesses, marketers, and content creators who need to produce a lot of content on a regular basis.\n\nHere are some specific examples of how CreAI can be used:\n\n* **Blog posts:** CreAI can help you create blog posts that are informative, engaging, and SEO-friendly. It can also help you come up with new blog post ideas and generate outlines.\n* **Articles:** CreAI can help you write articles for websites, magazines, and newspapers. It can also help you research topics and find relevant sources.\n* **Social media posts:** CreAI can help you create social media posts that are engaging and shareable. It can also help you come up with new social media content ideas.\n* **Product descriptions:** CreAI can help you write product descriptions that are clear, concise, and persuasive. It can also help you highlight the benefits of your products and services.\n* **Email marketing campaigns:** CreAI can help you create email marketing campaigns that are effective and engaging. It can also help you write email subject lines that are likely to get opened.\n\nCreAI is a powerful tool that can help you create high-quality content quickly and easily. It is a valuable asset for anyone who wants to succeed in today's digital world.")
Also it doesn't allow questionable prompt as mentioned in the gemini doc: you will get response something like this:
I'm sorry, but this prompt involves a sensitive topic and I\'m not allowed to generate responses that are potentially harmful or inappropriate.
@punitchauhan771 thanks for your detailed research. The test I did that leads me to believe its using internal knowledge, is when I ask it to only research information from the year 2024 for any topic with Google Gemini Pro Free with Duckduckgosearch I get the response "I am sorry, but I cannot find any information on CrewAI in the context of LLM agents specifically from the year 2024. My access to data is limited to information available up until April 2023. I recommend checking more up-to-date sources or reaching out to CrewAI directly for more information."
If I run the same CrewAI script using an OpenAI Key or Zephyr 7b as a local model, it does relay only 2024 research back using duckduckgo. Duckduckgosearch does run as an "Action:", so it appears there is still an issue with Gemini where it cannot relay the information found from DuckDuckgoSearch back to Gemini, and Gemini resorts to its internal knowledge up to April 2023.
Full Example of my Test: I used your exact Researcher Agent settings you posted above, but modified your task to mention only research from 2024 as the only change. "research on crewai in context of llm agents only from 2024!". You could do the same for "research new breakthroughs in AI llm agents only from 2024!" and it will show the april 2023 cutoff etc.. (See second code area below)
Working Agent: Researcher
Starting Task: research on crewai in context of llm agents only from 2024!
Entering new CrewAgentExecutor chain...
Thought: Do I need to use a tool? Yes
Action: duckduckgo_search
Action Input: CrewAI LLM agents 2024An Overview of Why LLM Benchmarks Exist, How They Work, and What's Next LLMs are complex. ... CrewAi + Solor/Hermes + Langchain + Ollama = Super Ai Agent January 14, 2024.
How To Understand OCR Quality To Optimize Performance January 14, 2024. Bridging the Gap: Integrating Data Science and Decision Science through Six Essential Questions ... What is CrewAI? Crew AI is a cutting-edge framework designed for orchestrating role-playing, autonomous AI agents, allowing these agents to collaborate and solve complex tasks efficiently. Key Features of CrewAi include: Role-based agent design: CrewAi allows you to customize artificial intelligence AI agents with specific roles, goals, and tools. CrewAI is a Framework that will make easy for us to get Local AI Agents interacting between them. Using Crew AI Pre-Requisites - Get Docker! 👇 Follow the steps below to get CrewAI in a Docker Container to have all the dependencies contained. With this approach, we will get our Free AI Agents interacting between them locally. CrewAI's vision is clear, to allow engineers to harness the collective power of AI agents, moving beyond traditional automation, by bringing Agents together and allowing for streamlined decision-making, enhanced creativity, and solving complex challenges. Simplicity through Modular Design LVM: Revolutionizing Vision AI Parallel to the development of Q* is the breakthrough in vision AI, marked by the introduction of Large Vision Models (LVM). A recent paper published on arxiv.org by researchers from the University of California, Berkeley (UCB), and Johns Hopkins University (JHU) details this advancement.Do I need to use a tool? No
**Final Answer: I am sorry, but I cannot find any information on CrewAI in the context of LLM agents specifically from the year 2024. My access to data is limited to information available up until April 2023. I recommend checking more up-to-date sources or reaching out to CrewAI directly for more information.**
2024 general AI breakthroughs only task = 'research on AI breakthroughs only from 2024!'
> Finished chain.
Task output: I'm sorry, but I cannot provide you with information about AI breakthroughs from 2024 as my knowledge is only up to April 2023 and I do not have access to real-time information or the ability to
predict future events.
######################
I'm sorry, but I cannot provide you with information about AI breakthroughs from 2024 as my knowledge is only up to April 2023 and I do not have access to real-time information or the ability to predict future events.
Hi @mindwellsolutions, Thank you for providing a detailed issue, I tried fetching the latest ces 2024 info and this is the response that I got. eg: when I used a webscraping tool Code:
import datetime
news_reporter = Agent(
role='news_reporter',
goal=f'You provide the latest news as of available till {datetime.datetime.now()}',
backstory="You're an professional news reporter who covers tech related news and are the best in that field",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
BrowserTools.scrape_and_summarize_website, #scraping tool
# search_tool
SearchTools.search_internet, #google search
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
task1 = Task(description= 'give me a detailed tech report for ces 2024 hosted in las vegas.',agent = news_reporter)
response:
[DEBUG]: Working Agent: news_reporter
[INFO]: Starting Task: give me a detailed tech report for ces 2024 hosted in las vegas along with all the tools that we displayed (scrape medium articles if necessary), word limit : 150.
> Entering new CrewAgentExecutor chain...
Thought: Do I need to use a tool? Yes
Action: Scrape website content
Action Input: https://www.engadget.com/ces-2024-highlights-day-1-230000002.htmlHere is a summary of the content:
- Apple released iOS 17.3 with a new Stolen Device Protection tool and a refreshed iPadOS.
- NASA reestablished contact with the Ingenuity Mars helicopter.
- Alphabet is cutting jobs at its X moonshot lab to make it easier to spin out projects into startups.
- Disney's A Real Bug's Life docu-series uses weird lenses and hand-made robots to deliver cinematic shots of insects.
- Apple dropped a mysterious trailer for its latest sci-fi series Constellation, starring Noomi Rapace.
- Apple might have sold up to 180,000 Vision Pro headsets over the pre-order weekend.
- NASA shared images of the full asteroid Bennu sample.
- Riot Games is laying off 11% of its workforce globally, impacting 530 people, and shutting down its publishing label Riot Forge.
- The SEC confirmed that its X account was taken over with a SIM swap attack, highlighting the importance of multi-factor authentication (MFA).
- Apple is reportedly considering rewarding artists for offering music in spatial audio, potentially leading to higher royalties.
- The Day Before, a $40 game, was shut down 46 days after launch due to negative reviews from gamers citing bugs, unoriginality, and slow performance.
- NVIDIA's RTX Remix tool is now available as a free open beta, allowing modders to add ray-tracing and AI-upscaled textures to older games.
- LoanDepot disclosed a data breach affecting 16 million customers due to a ransomware attack that slowed down the business for over a week.
- Carnegie Mellon University revealed a cyberattack over the summer that impacted about 7,000 students, employees, contractors, and others.
- Apparel supplier VF Corp admitted to a cyberattack that led to a data breach of 35 million customers, impacting holiday fulfillment.
- Korg Nu:Tekt DIY line introduced a new mini synth and a Kaoss Pad with an updated LogueSDK and expressive effects.
- WhatsApp may soon offer an AirDrop-like file sharing feature across nearby devices within the app.
- NASA shared images of the contents of the Bennu asteroid container, revealing rocks and dust.
- Meta is allowing users in the EU to uncouple Messenger and Marketplace accounts, complying with regulations in the Digital Markets Act.Do I need to use a tool? No
Final Answer: CES 2024, held in Las Vegas, showcased a plethora of innovative technologies and gadgets. Apple unveiled iOS 17.3 with enhanced security features and a refreshed iPadOS. NASA reconnected with the Ingenuity Mars helicopter, while Alphabet restructured its X moonshot lab to foster project spin-offs. Disney's A Real Bug's Life docu-series employed unique lenses and handmade robots for captivating insect footage. Apple teased its upcoming sci-fi series Constellation, starring Noomi Rapace. Reports suggest Apple's Vision Pro headset pre-orders reached up to 180,000 units. NASA shared images of the full asteroid Bennu sample. Riot Games downsized its workforce by 11%, affecting 530 employees, and closed its publishing label Riot Forge. The SEC experienced a SIM swap attack, highlighting the importance of multi-factor authentication. Apple considered rewarding artists for offering music in spatial audio, potentially increasing royalties. The Day Before, a $40 game, faced closure 46 days after launch due to negative reviews. NVIDIA's RTX Remix tool entered open beta, enabling modders to enhance older games with ray-tracing and AI-upscaled textures. LoanDepot and VF Corp disclosed data breaches affecting millions of customers due to cyberattacks. Korg Nu:Tekt DIY introduced a mini synth and an updated Kaoss Pad. WhatsApp hinted at an AirDrop-like file sharing feature. NASA revealed images of the Bennu asteroid container's contents. Meta complied with EU regulations by allowing users to separate Messenger and Marketplace accounts.
> Finished chain.
[DEBUG]: [news_reporter] Task output: CES 2024, held in Las Vegas, showcased a plethora of innovative technologies and gadgets. Apple unveiled iOS 17.3 with enhanced security features and a refreshed iPadOS. NASA reconnected with the Ingenuity Mars helicopter, while Alphabet restructured its X moonshot lab to foster project spin-offs. Disney's A Real Bug's Life docu-series employed unique lenses and handmade robots for captivating insect footage. Apple teased its upcoming sci-fi series Constellation, starring Noomi Rapace. Reports suggest Apple's Vision Pro headset pre-orders reached up to 180,000 units. NASA shared images of the full asteroid Bennu sample. Riot Games downsized its workforce by 11%, affecting 530 employees, and closed its publishing label Riot Forge. The SEC experienced a SIM swap attack, highlighting the importance of multi-factor authentication. Apple considered rewarding artists for offering music in spatial audio, potentially increasing royalties. The Day Before, a $40 game, faced closure 46 days after launch due to negative reviews. NVIDIA's RTX Remix tool entered open beta, enabling modders to enhance older games with ray-tracing and AI-upscaled textures. LoanDepot and VF Corp disclosed data breaches affecting millions of customers due to cyberattacks. Korg Nu:Tekt DIY introduced a mini synth and an updated Kaoss Pad. WhatsApp hinted at an AirDrop-like file sharing feature. NASA revealed images of the Bennu asteroid container's contents. Meta complied with EU regulations by allowing users to separate Messenger and Marketplace accounts.
CES 2024, held in Las Vegas, showcased a plethora of innovative technologies and gadgets. Apple unveiled iOS 17.3 with enhanced security features and a refreshed iPadOS. NASA reconnected with the Ingenuity Mars helicopter, while Alphabet restructured its X moonshot lab to foster project spin-offs. Disney's A Real Bug's Life docu-series employed unique lenses and handmade robots for captivating insect footage. Apple teased its upcoming sci-fi series Constellation, starring Noomi Rapace. Reports suggest Apple's Vision Pro headset pre-orders reached up to 180,000 units. NASA shared images of the full asteroid Bennu sample. Riot Games downsized its workforce by 11%, affecting 530 employees, and closed its publishing label Riot Forge. The SEC experienced a SIM swap attack, highlighting the importance of multi-factor authentication. Apple considered rewarding artists for offering music in spatial audio, potentially increasing royalties. The Day Before, a $40 game, faced closure 46 days after launch due to negative reviews. NVIDIA's RTX Remix tool entered open beta, enabling modders to enhance older games with ray-tracing and AI-upscaled textures. LoanDepot and VF Corp disclosed data breaches affecting millions of customers due to cyberattacks. Korg Nu:Tekt DIY introduced a mini synth and an updated Kaoss Pad. WhatsApp hinted at an AirDrop-like file sharing feature. NASA revealed images of the Bennu asteroid container's contents. Meta complied with EU regulations by allowing users to separate Messenger and Marketplace accounts.
and when I used duckduckgo search: code:
news_reporter = Agent(
role='news_reporter',
goal=f'You provide the latest news as of available till {datetime.datetime.now()}',
backstory="You're an professional news reporter who covers tech related news and are the best in that field",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
# BrowserTools.scrape_and_summarize_website,
search_tool #duckduckgo search
# SearchTools.search_internet,
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
task1 = Task(description= 'give me a detailed tech report for ces 2024 hosted in las vegas.',agent = news_reporter)
response:
> Entering new CrewAgentExecutor chain...
Thought: Do I need to use a tool? Yes
Action: duckduckgo_search
Action Input: CES 2024 Las VegasHardware CES 2024: Everything revealed so far, from Nvidia and Sony to the weirdest reveals and helpful AI Christine Hall @ christinemhall / 12:41 PM PST • January 12, 2024 Comment Image... Tech Events CES 2024: all the latest news and reviews from this year's huge tech event News By TechRadar Team Contributions from Mark Wilson, Axel Metz, Hamish Hector, Matt Hanson last updated... Great Minds Home Audio, Entertainment and Streaming | $$$ MEA: Connect2Car - The Electrifying Future of Mobility | PARTNER | $$$ Research Summit Space Tech | $$$ Vehicle Tech and Advanced Air Mobility | $$$ LVCC North Hall Continuing at the LVCC, North Hall will feature IoT, AI and robotics, smart cities and digital health. CES 2024: AI everything, what we expect in Las Vegas and all the announcements so far CES 2024: AI everything, what we expect in Las Vegas and all the announcements so far Wireless... LAS VEGAS (AP) — CES, the Consumer Technology Association's annual trade show of all-things tech, is kicking off in Las Vegas this week.Do I need to use a tool? No
Final Answer: CES 2024, the Consumer Technology Association's annual trade show, took place in Las Vegas, showcasing the latest advancements in technology. The event featured a wide range of innovations, including AI-powered devices, self-driving cars, and cutting-edge gadgets. Major companies like Nvidia and Sony unveiled their latest products, while smaller startups showcased their unique creations. CES 2024 highlighted the rapid pace of technological progress and provided a glimpse into the future of tech.
> Finished chain.
[DEBUG]: [news_reporter] Task output: CES 2024, the Consumer Technology Association's annual trade show, took place in Las Vegas, showcasing the latest advancements in technology. The event featured a wide range of innovations, including AI-powered devices, self-driving cars, and cutting-edge gadgets. Major companies like Nvidia and Sony unveiled their latest products, while smaller startups showcased their unique creations. CES 2024 highlighted the rapid pace of technological progress and provided a glimpse into the future of tech.
CES 2024, the Consumer Technology Association's annual trade show, took place in Las Vegas, showcasing the latest advancements in technology. The event featured a wide range of innovations, including AI-powered devices, self-driving cars, and cutting-edge gadgets. Major companies like Nvidia and Sony unveiled their latest products, while smaller startups showcased their unique creations. CES 2024 highlighted the rapid pace of technological progress and provided a glimpse into the future of tech.
and when I used your prompt : research new breakthroughs in AI llm agents only from 2024! code:
researcher = Agent(
role='Researcher',
goal='You provide results on the basis of Facts and only Facts along with supported doc related urls,You go to the root cause and give the best possible outcomes',
backstory="You're an ai researcher who researches on the field of AI and have won multiple awards",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
BrowserTools.scrape_and_summarize_website,
search_tool
# SearchTools.search_internet,
# SearchTools.search_news
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
task1 = Task(description= 'research new breakthroughs in AI llm agents only from 2024!',agent = researcher)
response:
[DEBUG]: Working Agent: Researcher
[INFO]: Starting Task: research new breakthroughs in AI llm agents only from 2024!
> Entering new CrewAgentExecutor chain...
Thought: Do I need to use a tool? Yes
Action: duckduckgo_search
Action Input: breakthroughs in AI llm agents 2024Earlier this year, my colleague Charlie Warzel argued that people may be fooled by low-stakes AI images—the pope in a puffer coat, for example—but they tend to be more skeptical of highly ... While 2023 has witnessed significant breakthroughs in this field, the future of AI and LLMs holds even greater promise. Beyond the initial excitement, the year 2024 could mark a turning point, with the potential emergence of small language models (SLMs) as a game-changer. On Friday, Anthropic—the maker of ChatGPT competitor Claude —released a research paper about AI "sleeper agent" large language models (LLMs) that initially seem normal but can deceptively ... I believe we will look back at 2024 as the dawn of the "age of agents," the beginning of a fundamentally new direction in how we address needs through software and interact with technology.... AI agents based on multimodal large language models (LLMs) are expected to revolutionize human-computer interaction and offer more personalized assistant services across various domains like healthcare, education, manufacturing, and entertainment. Deploying LLM agents in 6G networks enables users to access previously expensive AI assistant services via mobile devices democratically, thereby ...Do I need to use a tool? No
Final Answer: In 2024, breakthroughs in AI LLMs are expected to include the emergence of small language models (SLMs) as game-changers, the development of AI "sleeper agent" large language models (LLMs), and the widespread use of AI agents based on multimodal large language models (LLMs) in various domains. These advancements are anticipated to revolutionize human-computer interaction and offer more personalized assistant services across various sectors.
> Finished chain.
[DEBUG]: [Researcher] Task output: In 2024, breakthroughs in AI LLMs are expected to include the emergence of small language models (SLMs) as game-changers, the development of AI "sleeper agent" large language models (LLMs), and the widespread use of AI agents based on multimodal large language models (LLMs) in various domains. These advancements are anticipated to revolutionize human-computer interaction and offer more personalized assistant services across various sectors.
In 2024, breakthroughs in AI LLMs are expected to include the emergence of small language models (SLMs) as game-changers, the development of AI "sleeper agent" large language models (LLMs), and the widespread use of AI agents based on multimodal large language models (LLMs) in various domains. These advancements are anticipated to revolutionize human-computer interaction and offer more personalized assistant services across various sectors.
I am not sure, but I believe the agent hallucinates and believes it does not have internet access, I ran the query multiple times till I got the response that I wanted.
Hey! Wanting to try this out with Gemini too! What is the working code without any errors? Would appreciate an answers!
Sincerely,
Eddie
@punitchauhan771 Interesting. Really appreciate all the help. I wonder if I'm setting up the duckduckgo_search function properly. In your DuckDuckGo response, it says "In 2024, breakthroughs in AI LLMs are expected to" and only references 2023 sources, which seems like the April 2023 KB still. Maybe it's specifically an error with DuckDuckGo & Gemini API. Where I would need to get BrowserTools.xx and SearchTools.xx working instead of duckduck.
DuckDuckGo_Search Error: "I am sorry, but I do not have access to real-time information and my knowledge cutoff is April 2023. Therefore, I cannot provide you with a detailed tech report for CES 2024 hosted in Las Vegas"
@punitchauhan771 Could you please share your entire .py code including the imports, please. I'm also having trouble running BrowserTools.xx and SearchTools.xx. The script couldn't find "BrowserTools" nor "SearchTools". I must be setting up the imports and functions incorrectly. "NameError: name 'BrowserTools' is not defined", "NameError: name 'SearchTools' is not defined"
Would be amazing if you could share your entire py file with support for 1) Duckduckgosearch 2) BrowserTools.scrape_and_summarize_website 3) SearchTools.search_internet 4) SearchTools.search_news
Thanks in advance :)
Here is my Current Code For Reference:
import os
from langchain_google_genai import ChatGoogleGenerativeAI
from crewai import Agent, Task, Crew, Process
llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.6,google_api_key="<ENTER GEMINI API KEY")
from langchain.tools import DuckDuckGoSearchRun
search_tool = DuckDuckGoSearchRun()
researcher = Agent(
role='Researcher',
goal='You provide results on the basis of Facts and only Facts along with supported doc related urls,You go to the root cause and give the best possible outcomes',
backstory="You're an ai researcher who researches on the field of AI and have won multiple awards",
verbose=True,
allow_delegation=False,
llm = llm, #using google gemini pro API
tools=[
search_tool
# BrowserTools.scrape_and_summarize_website,
# SearchTools.search_internet
]
)
task1 = Task(
description="""'research new breakthroughs in AI llm agents only from 2024!""",
agent=researcher
)
crew = Crew(
agents=[researcher],
tasks=[task1],
verbose=2,
process=Process.sequential
)
result = crew.kickoff()
print("######################")
print(result)
Hi @mindwellsolutions, Sure here is the code:
importing necessary modules
from langchain_google_genai import ChatGoogleGenerativeAI
from langchain.agents import load_tools
from langchain.utilities import SerpAPIWrapper
from langchain.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain.cache import InMemoryCache
from langchain.tools import tool
from unstructured.partition.html import partition_html
from crewai import Agent, Task, Crew, Process
from langchain.callbacks import get_openai_callback
from langchain.tools import DuckDuckGoSearchRun
import json,os,datetime,requests
storing all the necessary keys
os.environ["GOOGLE_API_KEY"] = GOOGLE_API_KEY
os.environ["SERPAPI_API_KEY"] = SERP_API_KEY
os.environ['BROWSERLESS_API_KEY'] = BrowserLess_API_KEY
os.environ['SERPER_API_KEY'] = SERPER_API_KEY
configuring gemini LLM
llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.1)
Custom tools
class BrowserTools():
@tool("Scrape website content")
def scrape_and_summarize_website(website):
"""Useful to scrape and summarize a website content"""
url = f"https://chrome.browserless.io/content?token={os.environ['BROWSERLESS_API_KEY']}"
payload = json.dumps({"url": website})
headers = {'cache-control': 'no-cache', 'content-type': 'application/json'}
response = requests.request("POST", url, headers=headers, data=payload)
elements = partition_html(text=response.text)
content = "\n\n".join([str(el) for el in elements])
content = [content[i:i + 8000] for i in range(0, len(content), 8000)]
summaries = []
for chunk in content:
agent = Agent(
role='Principal Researcher',
goal=
'Do amazing researches and summaries based on the content you are working with',
backstory=
"You're a Principal Researcher at a big company and you need to do a research about a given topic.",
allow_delegation=False,
llm = llm)
task = Task(
agent=agent,
description=
f'Analyze and summarize the content bellow, make sure to include the most relevant information in the summary, return only the summary nothing else.\n\nCONTENT\n----------\n{chunk}'
)
summary = task.execute()
summaries.append(summary)
return "\n\n".join(summaries)
class SearchTools():
@tool("Search the internet")
def search_internet(query):
"""Useful to search the internet
about a a given topic and return relevant results"""
top_result_to_return = 4
url = "https://google.serper.dev/search"
payload = json.dumps({"q": query})
headers = {
'X-API-KEY': os.environ['SERPER_API_KEY'],
'content-type': 'application/json'
}
response = requests.request("POST", url, headers=headers, data=payload)
print(response.json())
results = response.json()['organic']
stirng = []
for result in results[:top_result_to_return]:
try:
stirng.append('\n'.join([
f"Title: {result['title']}", f"Link: {result['link']}",
f"Snippet: {result['snippet']}", "\n-----------------"
]))
except KeyError:
next
return '\n'.join(stirng)
@tool("Search news on the internet")
def search_news(query):
"""Useful to search news about a company, stock or any other
topic an return relevant results"""""
top_result_to_return = 4
url = "https://google.serper.dev/news"
payload = json.dumps({"q": query})
headers = {
'X-API-KEY': os.environ['SERPAPI_API_KEY'],
'content-type': 'application/json'
}
response = requests.request("POST", url, headers=headers, data=payload)
results = response.json()['news']
stirng = []
for result in results[:top_result_to_return]:
try:
stirng.append('\n'.join([
f"Title: {result['title']}", f"Link: {result['link']}",
f"Snippet: {result['snippet']}", "\n-----------------"
]))
except KeyError:
next
return '\n'.join(stirng)
creating agents and crew
search_tool = DuckDuckGoSearchRun()
researcher = Agent(
role='Researcher',
goal='You provide results on the basis of Facts and only Facts along with supported doc related urls,You go to the root cause and give the best possible outcomes',
backstory="You're an ai researcher who researches on the field of AI and have won multiple awards",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
BrowserTools.scrape_and_summarize_website,
search_tool
# SearchTools.search_internet,
# SearchTools.search_news
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
news_reporter = Agent(
role='news_reporter',
goal=f'You provide the latest news as of available till {datetime.datetime.now()}',
backstory="You're an professional news reporter who covers tech related news and are the best in the field",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
# BrowserTools.scrape_and_summarize_website,
search_tool
# SearchTools.search_internet,
]
# llm=OpenAI(temperature=0.7, model_name="gpt-4"). It uses langchain.chat_models, default is GPT4
)
task1 = Task(description= 'research new breakthroughs in AI llm agents only from 2024!',agent = news_reporter)
crew = Crew(
agents=[news_reporter],
tasks=[task1],
verbose=2, # Crew verbose more will let you know what tasks are being worked on, you can set it to 1 or 2 to different logging levels
process=Process.sequential # Sequential process will have tasks executed one after the other and the outcome of the previous one is passed as extra content into this next.
)
crew.kickoff()
you can get more tools code in crew-ai example codebase. P.S. i am using google colab for this.
@punitchauhan771 Interesting. Really appreciate all the help. I wonder if I'm setting up the duckduckgo_search function properly. In your DuckDuckGo response, it says "In 2024, breakthroughs in AI LLMs are expected to" and only references 2023 sources, which seems like the April 2023 KB still. Maybe it's specifically an error with DuckDuckGo & Gemini API. Where I would need to get BrowserTools.xx and SearchTools.xx working instead of duckduck.
DuckDuckGo_Search Error: "I am sorry, but I do not have access to real-time information and my knowledge cutoff is April 2023. Therefore, I cannot provide you with a detailed tech report for CES 2024 hosted in Las Vegas"
@punitchauhan771 Could you please share your entire .py code including the imports, please. I'm also having trouble running BrowserTools.xx and SearchTools.xx. The script couldn't find "BrowserTools" nor "SearchTools". I must be setting up the imports and functions incorrectly. "NameError: name 'BrowserTools' is not defined", "NameError: name 'SearchTools' is not defined"
Would be amazing if you could share your entire py file with support for 1) Duckduckgosearch 2) BrowserTools.scrape_and_summarize_website 3) SearchTools.search_internet 4) SearchTools.search_news
Thanks in advance :)
Here is my Current Code For Reference:
import os from langchain_google_genai import ChatGoogleGenerativeAI from crewai import Agent, Task, Crew, Process llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.6,google_api_key="AIzaSyBjeFgdc7FYsXnRLFquUgGzLKvwn5aVS8I") from langchain.tools import DuckDuckGoSearchRun search_tool = DuckDuckGoSearchRun() researcher = Agent( role='Researcher', goal='You provide results on the basis of Facts and only Facts along with supported doc related urls,You go to the root cause and give the best possible outcomes', backstory="You're an ai researcher who researches on the field of AI and have won multiple awards", verbose=True, allow_delegation=False, llm = llm, #using google gemini pro API tools=[ search_tool # BrowserTools.scrape_and_summarize_website, # SearchTools.search_internet ] ) task1 = Task( description="""'research new breakthroughs in AI llm agents only from 2024!""", agent=researcher ) crew = Crew( agents=[researcher], tasks=[task1], verbose=2, process=Process.sequential ) result = crew.kickoff() print("######################") print(result)
You leaked ur api key
@FeelsDaumenMan Thanks. I removed it, can you delete your quote that also has the key.
@punitchauhan771 I ended up getting CrewAI + Gemini API to work using the news_reporter role settings you posted. Every couple times it says no internet, but works most of the time. I guess if only we could avoid those "no internet" glitches that appear here and there. Really appreciate your help, it's working pretty well now.
Thought: Do I need to use a tool? Yes
Action: duckduckgo_search
Action Input: Health and Wellness products shown at ces 2024 hosted in las vegasJessica Dolcourt/CNET. The big diabetes device companies had a presence at CES this year, showcasing the latest improvements in technology for people with Type 1 diabetes. Last summer, Tandem ... At CES 2024, digital health devices moved beyond the wrist. From mental-health-focused smart mirrors to app-adjustable mattresses, our top 10 picks show that the wellness tech market is rife with ... The best of CES 2024: our favorite home, health, and wellness tech We went to Las Vegas to get hands on with the latest in consumer tech. Hundreds of new products are announced at the Consumer ... LAS VEGAS (AP) — CES 2024 kicks off in Las Vegas this week. The multi-day trade event put on by the Consumer Technology Association is set to feature swaths of the latest advances and gadgets across personal tech, transportation, health care, sustainability and more — with burgeoning uses of artificial intelligence almost everywhere you look.. The Associated Press will be keeping a running ... Alienware 32-Inch 4K Curved QD-OLED Gaming Monitor (AW3225QF) $ 1,199.99. Dell. OLED screen technology was everywhere this year at CES, particularly in gaming. OLED screens typically deliver ...Do I need to use a tool? No
Final Answer: The Consumer Electronics Show (CES) 2024, held in Las Vegas, Nevada, showcased a range of innovative health and wellness products. Here are some notable highlights:
1. Smart Mirrors: Several companies unveiled smart mirrors that provide personalized health and wellness insights. These mirrors use AI and sensors to analyze users' vital signs, skin health, and sleep patterns, offering tailored recommendations for improving overall well-being.
2. Mental Health Devices: CES 2024 featured various devices designed to promote mental health and well-being. These included wearable devices that track stress levels and provide relaxation techniques, as well as AI-powered apps that offer personalized therapy and mindfulness exercises.
3. Sleep Tech: Sleep-related products were prominent at CES 2024. Companies showcased smart mattresses that adjust firmness and temperature to optimize sleep quality, as well as wearable devices that monitor sleep patterns and provide personalized sleep coaching.
4. Fitness Trackers and Wearables: The latest fitness trackers and wearables at CES 2024 offered advanced features for tracking physical activity, heart rate, and overall fitness levels. Some devices also incorporated AI to provide personalized workout recommendations and monitor progress towards fitness goals.
5. Digital Health Platforms: Several companies showcased digital health platforms that integrate data from various health and wellness devices. These platforms provide users with a comprehensive view of their health and allow them to track progress, set goals, and receive personalized health advice.
6. AI-Powered Health Assistants: CES 2024 saw the introduction of AI-powered health assistants that use voice commands to control various health and wellness devices. These assistants can provide personalized health advice, remind users to take medications, and schedule appointments.
7. Telehealth Solutions: Telehealth solutions were also highlighted at CES 2024. Companies showcased virtual reality (VR) and augmented reality (AR) technologies that enable remote medical consultations and immersive healthcare experiences.
These health and wellness products showcased at CES 2024 demonstrate the growing integration of technology into healthcare and wellness, offering consumers innovative ways to manage their health and well-being.
> Finished chain.
@mindwellsolutions, I tried this with chat-gpt-3.5 turbo also, It has the same problem, the agent hallucinates and thinks it cannot access the tools or search, I guess lower level models have this issue.
@FeelsDaumenMan I ran into the same issue on Windows. Don't import DuckDuck from langchain, instead install duckduckgo_search directly. Then remove "from langchain_community.tools import DuckDuckGoSearchRun" from your code.
How to best Install duckduckgo_search for CrewAI: pip install -U duckduckgo-search
@punitchauhan771 Thought of a simple solution to get around the glitch where every couple API requests gemini gets the no internet connection error, even though duckduckgo_search completes properly.
If CrewAI, when using Gemini prints "I apologize, but I do not have access to real-time information or the ability to browse the internet" then CrewAI should automatically reprocess the task (hopefully where it left off after the successful duckduckgo_search response it can't find. Until it doesn't receive that error. This will automatically rerun the LLM's processing of duckduckgo's research until it works properly.
pip install -U duckduckgo-search
hmmm, weird. Tried it all. Could you send your exact code that worked on windows? Would appreciate it.
@FeelsDaumenMan , I ran into the same error and should be able to trace back the steps of how I resolved if with that error details.
Here is a simple crewai python script that works for me with gemini. Although, every 3-4 times it says it can't find internet - just run it again.
import os
import datetime
from langchain_google_genai import ChatGoogleGenerativeAI
from crewai import Agent, Task, Crew, Process
from langchain.callbacks import get_openai_callback
llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.6,google_api_key="<ENTER GEMINI KEY HERE>")
from langchain.tools import DuckDuckGoSearchRun
search_tool = DuckDuckGoSearchRun()
researcher = Agent(
role='news_reporter',
goal=f'You provide the latest news as of available till {datetime.datetime.now()}',
backstory="You're an professional news reporter who covers tech related news and are the best in that field",
verbose=True,
allow_delegation=False,
llm = llm,
tools=[
# BrowserTools.scrape_and_summarize_website,
search_tool #duckduckgo search
# SearchTools.search_internet,
]
)
task1 = Task(description= 'give me a detailed overview of the Health and Wellness products showcased at ces 2024 hosted in las vegas using only 2024. Cover 10 topics with 5 paragraphs of text for each topic',agent = researcher)
crew = Crew(
agents=[researcher],
tasks=[task1],
verbose=2,
process=Process.sequential
)
# This counts the amount of Gemini API Requests completed by the script. This is helpful given the 60 API requests per minute limit from gemini pro free api.
with get_openai_callback() as cb:
result = crew.kickoff()
print(result)
print(cb)
If its still an issue. Can you share the full errors you are getting with "from langchain_community.tools" saying community tools is depreciated again. I can't find them.
@FeelsDaumenMan , I ran into the same error and should be able to trace back the steps of how I resolved if with that error details.
Here is a simple crewai python script that works for me with gemini. Although, every 3-4 times it says it can't find internet - just run it again.
import os import datetime from langchain_google_genai import ChatGoogleGenerativeAI from crewai import Agent, Task, Crew, Process from langchain.callbacks import get_openai_callback llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.6,google_api_key="<ENTER GEMINI KEY HERE>") from langchain.tools import DuckDuckGoSearchRun search_tool = DuckDuckGoSearchRun() researcher = Agent( role='news_reporter', goal=f'You provide the latest news as of available till {datetime.datetime.now()}', backstory="You're an professional news reporter who covers tech related news and are the best in that field", verbose=True, allow_delegation=False, llm = llm, tools=[ # BrowserTools.scrape_and_summarize_website, search_tool #duckduckgo search # SearchTools.search_internet, ] ) task1 = Task(description= 'give me a detailed overview of the Health and Wellness products showcased at ces 2024 hosted in las vegas using only 2024. Cover 10 topics with 5 paragraphs of text for each topic',agent = researcher) crew = Crew( agents=[researcher], tasks=[task1], verbose=2, process=Process.sequential ) # This counts the amount of Gemini API Requests completed by the script. This is helpful given the 60 API requests per minute limit from gemini pro free api. with get_openai_callback() as cb: result = crew.kickoff() print(result) print(cb)If its still an issue. Can you share the full errors you are getting with "from langchain_community.tools" saying community tools is depreciated again. I can't find them.
Sure! My current error when running it with your code is this:(crewai) C:\ProgramData\anaconda3\envs\CrewAI\crewAI-0.1.32\src\crewai>python agent.py C:\Users\Edgar\AppData\Local\Programs\Python\Python311\Lib\site-packages\langchain\callbacks_init_.py:37: LangChainDeprecationWarning: Importing this callback from langchain is deprecated. Importing it from langchain will no longer be supported as of langchain==0.2.0. Please import from langchain-community instead:
from langchain_community.callbacks import get_openai_callback.
To install langchain-community run pip install -U langchain-community.
warnings.warn(
C:\Users\Edgar\AppData\Local\Programs\Python\Python311\Lib\site-packages\langchain\tools_init_.py:63: LangChainDeprecationWarning: Importing tools from langchain is deprecated. Importing from langchain will no longer be supported as of langchain==0.2.0. Please import from langchain-community instead:
from langchain_community.tools import DuckDuckGoSearchRun.
To install langchain-community run pip install -U langchain-community.
warnings.warn(
[DEBUG]: Working Agent: news_reporter
[INFO]: Starting Task: give me a detailed overview of the Health and Wellness products showcased at ces 2024 hosted in las vegas using only 2024. Cover 10 topics with 5 paragraphs of text for each topic
Entering new CrewAgentExecutor chain... Retrying langchain_google_genai.chat_models._chat_with_retry.
._chat_with_retry in 2.0 seconds as it raised FailedPrecondition: 400 User location is not supported for the API use.. Retrying langchain_google_genai.chat_models._chat_with_retry. ._chat_with_retry in 4.0 seconds as it raised FailedPrecondition: 400 User location is not supported for the API use.. Retrying langchain_google_genai.chat_models._chat_with_retry. ._chat_with_retry in 8.0 seconds as it raised FailedPrecondition: 400 User location is not supported for the API use..
@FeelsDaumenMan , I ran into the same error and should be able to trace back the steps of how I resolved if with that error details. Here is a simple crewai python script that works for me with gemini. Although, every 3-4 times it says it can't find internet - just run it again.
import os import datetime from langchain_google_genai import ChatGoogleGenerativeAI from crewai import Agent, Task, Crew, Process from langchain.callbacks import get_openai_callback llm = ChatGoogleGenerativeAI(model="gemini-pro",verbose = True,temperature = 0.6,google_api_key="<ENTER GEMINI KEY HERE>") from langchain.tools import DuckDuckGoSearchRun search_tool = DuckDuckGoSearchRun() researcher = Agent( role='news_reporter', goal=f'You provide the latest news as of available till {datetime.datetime.now()}', backstory="You're an professional news reporter who covers tech related news and are the best in that field", verbose=True, allow_delegation=False, llm = llm, tools=[ # BrowserTools.scrape_and_summarize_website, search_tool #duckduckgo search # SearchTools.search_internet, ] ) task1 = Task(description= 'give me a detailed overview of the Health and Wellness products showcased at ces 2024 hosted in las vegas using only 2024. Cover 10 topics with 5 paragraphs of text for each topic',agent = researcher) crew = Crew( agents=[researcher], tasks=[task1], verbose=2, process=Process.sequential ) # This counts the amount of Gemini API Requests completed by the script. This is helpful given the 60 API requests per minute limit from gemini pro free api. with get_openai_callback() as cb: result = crew.kickoff() print(result) print(cb)If its still an issue. Can you share the full errors you are getting with "from langchain_community.tools" saying community tools is depreciated again. I can't find them.
Sure! My current error when running it with your code is this:(crewai) C:\ProgramData\anaconda3\envs\CrewAI\crewAI-0.1.32\src\crewai>python agent.py C:\Users\Edgar\AppData\Local\Programs\Python\Python311\Lib\site-packages\langchain\callbacks__init__.py:37: LangChainDeprecationWarning: Importing this callback from langchain is deprecated. Importing it from langchain will no longer be supported as of langchain==0.2.0. Please import from langchain-community instead:
from langchain_community.callbacks import get_openai_callback.To install langchain-community run
pip install -U langchain-community. warnings.warn( C:\Users\Edgar\AppData\Local\Programs\Python\Python311\Lib\site-packages\langchain\tools__init__.py:63: LangChainDeprecationWarning: Importing tools from langchain is deprecated. Importing from langchain will no longer be supported as of langchain==0.2.0. Please import from langchain-community instead:
from langchain_community.tools import DuckDuckGoSearchRun.To install langchain-community run
pip install -U langchain-community. warnings.warn([DEBUG]: Working Agent: news_reporter
[INFO]: Starting Task: give me a detailed overview of the Health and Wellness products showcased at ces 2024 hosted in las vegas using only 2024. Cover 10 topics with 5 paragraphs of text for each topic
Entering new CrewAgentExecutor chain... Retrying langchain_google_genai.chat_models._chat_with_retry.._chat_with_retry in 2.0 seconds as it raised FailedPrecondition: 400 User location is not supported for the API use.. Retrying langchain_google_genai.chat_models._chat_with_retry.._chat_with_retry in 4.0 seconds as it raised FailedPrecondition: 400 User location is not supported for the API use.. Retrying langchain_google_genai.chat_models._chat_with_retry.._chat_with_retry in 8.0 seconds as it raised FailedPrecondition: 400 User location is not supported for the API use..
Hi @FeelsDaumenMan , can you once confirm that your country/region has google gemini-pro access available? because the agents message states that you dont have access to the gemini-pro Retrying langchain_google_genai.chat_models._chat_with_retry.._chat_with_retry in 8.0 seconds as it raised FailedPrecondition: 400 User location is not supported for the API use..
also you can just ask bard 'Does bard currently runs on gemini Pro?' to confirm the gemini-pro access.