LlamaResearcher: AI Research Assistant Based on Llama 4 That Turns Topics into Full Articles 132

What is LlamaResearcher?

LlamaResearcher is a research assistant built on the Llama 4 model that can quickly transform user-input topics into complete papers, suitable for a variety of scenarios including academic research, report writing, and content creation.

LlamaResearcher Features

1. Rapid Paper Generation

• Core Feature: Capable of quickly transforming user-input topics into fully structured papers, saving time on research and writing.

2. Security Check Mechanism

• Security Evaluation: Uses the llama-3-8b-guard model provided by Groq to perform security checks on user-input query requests, ensuring that the input content meets safety standards.

• Risk Control: Avoids processing requests that may involve inappropriate content, safeguarding system security and compliance.

3. Intelligent Query Expansion

• Multi-Dimensional Search: Expands a user's single query into three sub-queries, exploring the topic in depth from multiple angles.

• In-Depth Information Retrieval: Performs deep web searches for each sub-query through LinkUp, obtaining more comprehensive and richer information resources.

4. Information Filtering and Evaluation

• Relevance Evaluation: Filters the large amount of retrieved information to ensure that the returned content is highly relevant to the user's original query.

• Quality Control: Filters out irrelevant or low-quality information through an evaluation mechanism, ensuring that the final generated paper content is accurate and reliable.

5. Powerful Tech Stack Support

• Groq: Used for security checks and model inference.

• LinkUp: Used for deep web search.

• Llamaindex: Used for information retrieval and generation.

• Gradio: Provides a user-friendly interface.

• FastAPI: Builds efficient backend services.

• Redis: Used for API rate limiting, ensuring stable system operation.

LlamaResearcher Workflow

• Security Check: Uses the llama-3-8b-guard model provided by Groq to perform security evaluation on user requests.

• Task Assignment: If the request is safe, it is routed to the ResearcherAgent.

• Query Expansion: The ResearcherAgent expands the query into three sub-queries for web searching.

• Information Retrieval and Filtering: Uses LinkUp to perform deep web searches for each sub-query, and evaluates the relevance of the retrieved information to the user's original query.

• Paper Generation: After all information is collected, the ResearcherAgent writes the final paper and returns it to the user.

LlamaResearcher Installation and Running

1. Clone the repository:

git clone https://github.com/AstraBert/llama-4-researcher.git
cd llama-4-researcher

2. Configure environment variables:

• Rename the .env.example file to .env, and add the following API keys:

• groq_api_key: Obtained from the Groq Console.

• linkup_api_key: Obtained from the Linkup Dashboard.

• internal_api_key: Create it yourself.

• Or start manually:

docker compose up -f compose.local.yaml llama_redis -d
docker compose up -f compose.local.yaml llama_app -d

3. The application will run at http://localhost:8000.

Github address:https://github.com/AstraBert/llama-4-researcher