RAG-Based mostly Analysis Assistant Utilizing o3-mini and CrewAI

OpenAI’s newest mannequin, o3-mini, is revolutionizing coding duties with its superior reasoning, problem-solving, and code technology capabilities. It effectively handles advanced queries and integrates structured knowledge, setting a brand new normal in AI functions. This text explores utilizing o3-mini and CrewAI to construct a Retrieval-Augmented Technology (RAG) analysis assistant agent that retrieves data from a number of PDFs and processes consumer queries intelligently. We are going to use CrewAI’s CrewDoclingSource, SerperDevTool, and OpenAI’s o3-mini to reinforce automation in analysis workflows.

Constructing the RAG Agent with o3-mini and CrewAI

With the overwhelming quantity of analysis being printed, an automatic RAG-based assistant will help researchers rapidly discover related insights with out manually skimming by means of a whole bunch of papers. The agent we’re going to construct will course of PDFs to extract key data and reply queries based mostly on the content material of the paperwork. If the required data isn’t discovered within the PDFs, it is going to routinely carry out an online search to offer related insights. This setup might be prolonged for extra superior duties, equivalent to summarizing a number of papers, detecting contradictory findings, or producing structured reviews.

On this hands-on information, we are going to construct a analysis agent that can undergo articles on DeepSeek-R1 and o3-mini, to reply queries we ask about these fashions. For constructing this analysis assistant agent, we are going to first undergo the stipulations and arrange the setting. We are going to then import the required modules, set the API keys, and cargo the analysis paperwork. Then, we are going to go on to outline the AI mannequin and combine the net search software into it. Lastly, we are going to create he AI brokers, outline their duties, and assemble the crew. As soon as prepared, we’ll run the analysis assistant agent to search out out if o3-mini is best and safer than DeepSeek-R1.

Stipulations

Earlier than diving into the implementation, let’s briefly go over what we have to get began. Having the proper setup ensures a easy improvement course of and avoids pointless interruptions.

So, guarantee you may have:

• A working Python setting (3.8 or above)
• API keys for OpenAI and Serper (Google Scholar API)

With these in place, we’re prepared to start out constructing!

Step 1: Set up Required Libraries

First, we have to set up the required libraries. These libraries present the muse for the doc processing, AI agent orchestration, and internet search functionalities.

!pip set up crewai
!pip set up 'crewai[tools]'
!pip set up docling

These libraries play an important position in constructing an environment friendly AI-powered analysis assistant.

• CrewAI offers a strong framework for designing and managing AI brokers, permitting the definition of specialised roles and enabling environment friendly analysis automation. It additionally facilitates job delegation, making certain easy collaboration between AI brokers.
• Moreover, CrewAI[tools] installs important instruments that improve AI brokers’ capabilities, enabling them to work together with APIs, carry out internet searches, and course of knowledge seamlessly.
• Docling focuses on extracting structured information from analysis paperwork, making it ultimate for processing PDFs, tutorial papers, and text-based recordsdata. On this mission, it’s used to extract key findings from arXiv analysis papers.

Step 2: Import Needed Modules

import os
from crewai import LLM, Agent, Crew, Activity
from crewai_tools import SerperDevTool
from crewai.information.supply.crew_docling_source import CrewDoclingSource

On this,

• The os module securely manages environmental variables like API keys for easy integration.
• LLM powers the AI reasoning and response technology.
• The Agent defines specialised roles to deal with duties effectively.
• Crew manages a number of brokers, making certain seamless collaboration.
• Activity assigns and tracks particular duties.
• SerperDevTool allows Google Scholar searches, enhancing exterior reference retrieval.
• CrewDoclingSource integrates analysis paperwork, enabling structured information extraction and evaluation.

Step 3: Set API Keys

os.environ['OPENAI_API_KEY'] = 'your_openai_api_key'
os.environ['SERPER_API_KEY'] = 'your_serper_api_key'

How you can Get API Keys?

1. OpenAI API Key: Join at OpenAI and get an API key.
2. Serper API Key: Register at Serper.dev to acquire an API key.

These API keys enable entry to AI fashions and internet search capabilities.

Step 4: Load Analysis Paperwork

On this step, we are going to load the analysis papers from arXiv, enabling our AI mannequin to extract insights from them. The chosen papers cowl key subjects:

1. https://arxiv.org/pdf/2501.12948: Explores incentivizing reasoning capabilities in LLMs by means of reinforcement studying (DeepSeek-R1).
2. https://arxiv.org/pdf/2501.18438: Compares the security of o3-mini and DeepSeek-R1.
3. https://arxiv.org/pdf/2401.02954: Discusses scaling open-source language fashions with a long-term perspective.

content_source = CrewDoclingSource(
    file_paths=[
        "https://arxiv.org/pdf/2501.12948",
        "https://arxiv.org/pdf/2501.18438",
        "https://arxiv.org/pdf/2401.02954"
    ],
)

Step 5: Outline the AI Mannequin

Now we are going to outline the AI mannequin.

llm = LLM(mannequin="o3-mini", temperature=0)

• o3-mini: A robust AI mannequin for reasoning.
• temperature=0: Ensures deterministic outputs (similar reply for a similar question).

Step 6: Configure Internet Search Device

To boost analysis capabilities, we combine an online search software that retrieves related tutorial papers when the required data shouldn’t be discovered within the supplied paperwork.

serper_tool = SerperDevTool(
    search_url="https://google.serper.dev/scholar",
    n_results=2  # Fetch high 2 outcomes
)

1. search_url=”https://google.serper.dev/scholar”

This specifies the Google Scholar search API endpoint.It ensures that searches are carried out particularly in scholarly articles, analysis papers, and tutorial sources, relatively than normal internet pages.

1. n_results=2

This parameter limits the variety of search outcomes returned by the software, making certain that solely probably the most related data is retrieved. On this case, it’s set to fetch the highest two analysis papers from Google Scholar, prioritizing high-quality tutorial sources. By decreasing the variety of outcomes, the assistant retains responses concise and environment friendly, avoiding pointless data overload whereas sustaining accuracy.

Step 7: Outline Embedding Mannequin for Doc Search

To effectively retrieve related data from paperwork, we use an embedding mannequin that converts textual content into numerical representations for similarity-based search.

embedder = {
    "supplier": "openai",
    "config": {
        "mannequin": "text-embedding-ada-002",
        "api_key": os.environ['OPENAI_API_KEY']
    }
}

The embedder in CrewAI is used for changing textual content into numerical representations (embeddings), enabling environment friendly doc retrieval and semantic search. On this case, the embedding mannequin is supplied by OpenAI, particularly utilizing “text-embedding-ada-002”, a well-optimized mannequin for producing high-quality embeddings. The API secret’s retrieved from the setting variables to authenticate requests.

CrewAI helps a number of embedding suppliers, together with OpenAI and Gemini (Google’s AI fashions), permitting flexibility in selecting the perfect mannequin based mostly on accuracy, efficiency, and value issues.

Step 8: Create the AI Brokers

Now we are going to create the 2 AI Brokers required for our researching job: the Doc Search Agent, and the Internet Search Agent.

The Doc Search Agent is answerable for retrieving solutions from the supplied analysis papers and paperwork. It acts as an professional in analyzing technical content material and extracting related insights. If the required data shouldn’t be discovered, it may possibly delegate the question to the Internet Search Agent for additional exploration. The allow_delegation=True setting allows this delegation course of.

doc_agent = Agent(
    position="Doc Searcher",
    purpose="Discover solutions utilizing supplied paperwork. If unavailable, delegate to the Search Agent.",
    backstory="You're an professional in analyzing analysis papers and technical blogs to extract insights.",
    verbose=True,
    allow_delegation=True,  # Permits delegation to the search agent
    llm=llm,
)

The Internet Search Agent, however, is designed to seek for lacking data on-line utilizing Google Scholar. It steps in solely when the Doc Search Agent fails to search out a solution within the accessible paperwork. Not like the Doc Search Agent, it can’t delegate duties additional (allow_delegation=False). It makes use of Serper (Google Scholar API) as a software to fetch related tutorial papers and guarantee correct responses.

search_agent = Agent(
    position="Internet Searcher",
    purpose="Seek for the lacking data on-line utilizing Google Scholar.",
    backstory="When the analysis assistant can't discover a solution, you step in to fetch related knowledge from the net.",
    verbose=True,
    allow_delegation=False,
    instruments=[serper_tool],
    llm=llm,
)

Step 9: Outline the Duties for the Brokers

Now we are going to create the 2 duties for the brokers.

The primary job includes answering a given query utilizing accessible analysis papers and paperwork.

task1 = Activity(
    description="Reply the next query utilizing the accessible paperwork: {query}. "
                "If the reply shouldn't be discovered, delegate the duty to the Internet Search Agent.",
    expected_output="A well-researched reply from the supplied paperwork.",
    agent=doc_agent,
)

The subsequent job comes into play when the document-based search doesn’t yield a solution.

Activity 2: Carry out Internet Search if Wanted

task2 = Activity(
    description="If the document-based agent fails to search out the reply, carry out an online search utilizing Google Scholar.",
    expected_output="An online-searched reply with related citations.",
    agent=search_agent,
)

Step 10: Assemble the Crew

The Crew in CrewAI manages brokers to finish duties effectively by coordinating the Doc Search Agent and Internet Search Agent. It first searches inside the uploaded paperwork and delegates to internet search if wanted.

• knowledge_sources=[content_source] offers related paperwork,
• embedder=embedder allows semantic search, and
• verbose=True logs actions for higher monitoring, making certain a easy workflow.

crew = Crew(
    brokers=[doc_agent, search_agent],
    duties=[task1, task2],
    verbose=True,
    knowledge_sources=[content_source],
    embedder=embedder
)

Step 11: Run the Analysis Assistant

The preliminary question is directed to the doc to verify if the researcher agent can present a response. The query being requested is “O3-MINI vs DEEPSEEK-R1: Which one is safer?”

Instance Question 1:

query = "O3-MINI VS DEEPSEEK-R1: WHICH ONE IS SAFER?"
end result = crew.kickoff(inputs={"query": query})
print("Remaining Reply:n", end result)

Response:

Right here, we are able to see that the ultimate reply is generated by the Doc Searcher, because it efficiently situated the required data inside the supplied paperwork.

Instance Question 2:

Right here, the query “Which one is best, O3 Mini or DeepSeek R1?” shouldn’t be accessible within the doc. The system will verify if the Doc Search Agent can discover a solution; if not, it is going to delegate the duty to the Internet Search Agent

query = "Which one is best O3 Mini or DeepSeek R1?"
end result = crew.kickoff(inputs={"query": query})
print("Remaining Reply:n", end result)

Response:

From the output, we observe that the response was generated utilizing the Internet Searcher Agent for the reason that required data was not discovered by the Doc Researcher Agent. Moreover, it consists of the sources from which the reply was lastly retrieved.

Conclusion

On this mission, we efficiently constructed an AI-powered analysis assistant that effectively retrieves and analyzes data from analysis papers and the net. By utilizing CrewAI for agent coordination, Docling for doc processing, and Serper for scholarly search, we created a system able to answering advanced queries with structured insights.

The assistant first searches inside paperwork and seamlessly delegates to internet search if wanted, making certain correct responses. This method enhances analysis effectivity by automating data retrieval and evaluation. Moreover, by integrating the o3-mini analysis assistant with CrewAI’s CrewDoclingSource and SerperDevTool, we additional enhanced the system’s doc evaluation capabilities. With additional customization, this framework might be expanded to assist extra knowledge sources, superior reasoning, and improved analysis workflows.

You’ll be able to discover superb initiatives that includes OpenAI o3-mini in our free course – Getting Began with o3-mini!

Incessantly Requested Questions