LangGraph is a library for building stateful, multi-actor applications with LLMs. This example uses Langgraph to automate the process of an in depth research on any given topic.

By using Langgraph, the research process can be significantly improved in depth and quality by leveraging multiple agents with specialized skills. Inspired by the recent STORM paper, this example showcases how a team of AI agents can work together to conduct research on a given topic, from planning to publication.

An average run generates a 5-6 page research report in multiple formats such as PDF, Docx and Markdown.

The research team is made up of 7 AI agents:

• Chief Editor – Oversees the research process and manages the team. This is the “master” agent that coordinates the other agents using Langgraph.
• Researcher (gpt-researcher) – A specialized autonomous agent that conducts in depth research on a given topic.
• Editor – Responsible for planning the research outline and structure.
• Reviewer – Validates the correctness of the research results given a set of criteria.
• Revisor – Revises the research results based on the feedback from the reviewer.
• Writer – Responsible for compiling and writing the final report.
• Publisher – Responsible for publishing the final report in various formats.

Generally, the process is based on the following stages:

1. Planning stage
2. Data collection and analysis
3. Review and revision
4. Writing and submission
5. Publication

More specifically (as seen in the architecture diagram) the process is as follows:

• Browser (gpt-researcher) – Browses the internet for initial research based on the given research task.
• Editor – Plans the report outline and structure based on the initial research.
• For each outline topic (in parallel):
  
  • Researcher (gpt-researcher) – Runs an in depth research on the subtopics and writes a draft.
  • Reviewer – Validates the correctness of the draft given a set of criteria and provides feedback.
  • Revisor – Revises the draft until it is satisfactory based on the reviewer feedback.
• Writer – Compiles and writes the final report including an introduction, conclusion and references section from the given research findings.
• Publisher – Publishes the final report to multi formats such as PDF, Docx, Markdown, etc.

1. Install required packages:
   

   pip install -r requirements.txt

    
2. Update env variables
   

   export OPENAI_API_KEY={Your OpenAI API Key here}
   export TAVILY_API_KEY={Your Tavily API Key here}

    
3. Run the application:
   

   python main.py

    

To change the research query and customize the report, edit the task.json file in the main directory.

• query – The research query or task.
• model – The OpenAI LLM to use for the agents.
• max_sections – The maximum number of sections in the report. Each section is a subtopic of the research query.
• publish_formats – The formats to publish the report in. The reports will be written in the output directory.
• follow_guidelines – If true, the research report will follow the guidelines below. It will take longer to complete. If false, the report will be generated faster but may not follow the guidelines.
• guidelines – A list of guidelines that the report must follow.
• verbose – If true, the application will print detailed logs to the console.

{
  "query": "Is AI in a hype cycle?",
  "model": "gpt-4o",
  "max_sections": 3, 
  "publish_formats": { 
    "markdown": true,
    "pdf": true,
    "docx": true
  },
  "follow_guidelines": true,
  "guidelines": [
    "The report MUST fully answer the original question",
    "The report MUST be written in apa format",
    "The report MUST be written in english"
  ],
  "verbose": true
}