PubMed-Tool

You can finally trust AI with your research thanks to a direct and up-to-date link to PubMed. With the help of AI Agents, you can effortlessly chat with over 37 million biomedical citations. The Agent searches the PubMed database to match your query and provides you with concise responses.

What is the PubMed Tool

The PubMed Tool component lets you chat with millions of reputable biomedical articles. This way, you can turn hours of looking for the right sources into a quick chat. Based on your choices, it can either return articles from the PubMed database, along with a short description and a link, or piece together information from various entries.

Remember that PubMed is a database that links to the articles but doesn’t contain their full versions. That’s why FlowHunt can only access publicly available information on PubMed, such as the abstracts and authors.

Key Features

• Direct Access to PubMed: Seamlessly integrates PubMed search capabilities into your AI workflows.
• Biomedical & Medical Focus: Ideal for retrieving scientific articles, clinical studies, and up-to-date medical information.
• Flexible Tool Integration: Can be referenced and utilized as a tool by AI agents within larger workflows.

Inputs

While the component’s JSON definition includes some configurable fields, most are advanced and intended for internal or agent use:

These inputs are generally set internally and are not required for basic use. They provide flexibility for advanced users customizing agent behaviors or debugging.

Outputs

The main output of the PubMed component is:

Usage Scenarios

• Medical Q&A Agents: Enhance chatbots or virtual assistants with the ability to source answers from medical literature.
• Literature Review Automation: Automate the retrieval and summarization of recent research for clinicians or researchers.
• Health Information Systems: Integrate up-to-date PubMed data into health informatics platforms.

How to use the PubMed tool

This component works with an autonomous AI agent. AI Agents are computer programs that can perform tasks and make decisions independently. The Agents observe their environment and take action based on their unique programming, knowledge, and goals.

If you need more information, simply ask the bot. It’s just like chatting with a knowledgeable friend. It will get deeper into the topic, combining various articles or seamlessly switching to other related topics.

How to connect the Pubmed Tool to your Flow

Let’s create a basic PubMed Chabot. Thanks to the use of an AI Agent, the Flow is super simple and contains just a few components:

Steps:

• Add Chat Input and AI Agent.
• Connect the input to the Agent.
• Connect the PubMed Tool to the AI Agent’s Tool handle.
• Connect Chat History to let the past messages inform future answers. This is optional and costs more tokens, but it is necessary if you want to give follow-up questions.
• You’re ready to send the Agent’s text output to the chat.

Feel free to experiment and expand the Flow. You can add more search tools and Knowledge Retrievers to make the bot a true information powerhouse. Play around with the Agent’s goal and backstory to achieve your desired output.

Note: If you want the Chatbot to link sources, simply add this requirement to the Agent’s goal.

PubMed Tool in action

Let’s say you’re trying to find out what’s new in brain cancer research:

The Chatbot will return a couple of recently trending articles on the topic. You can specify how many you want in the chat or simply ask it to give you more options. To get more information about an entry, simply ask about it or follow the link.

What happens if there isn’t a PubMed entry that matches your query? The Chatbot lets you know about it instead of hallucinating an answer. Let’s test it. Let’s ask the PubMed Chatbot about an SEO SaaS it shouldn’t know anything about.

When the Agent can’t find any suitable PubMed articles, it apologizes and prompts you to search the web for your query instead: