ThoughtChain

ThoughtChain is an AI-powered assistant that builds a short literature review for a given research question. It queries Semantic Scholar, downloads open access papers, screens them with a language model and returns a concise scientific report. A small Flask web interface is also provided.

Installation

1. Install Python 3.8 or newer.
2. Install dependencies:

pip install -r requirements.txt
# additional packages for the web app
pip install flask flask-cors python-dotenv

Configuration

Create a .env file in the project root and add your API keys:

OPENAI_API_KEY=your-openai-key
SEMANTIC_SCHOLAR_API_KEY=your-semantic-scholar-key

The application loads this file automatically. Keep the .env file private and never commit it to version control.

Usage

Command line example

You can call the pipeline directly from Python:

python - <<'PY'
from main import run_pipeline
result = run_pipeline("What is the role of epigenetics in Alzheimer's disease?")
print(result["report"])
PY

Web interface

To start the Flask interface run:

python app.py

Open http://127.0.0.1:5000/ in your browser and enter a research question. The page will display each step of the pipeline and the generated report.

Further Development

Want to build on this project? Here are a few ideas:

• Add PDF parsing: Allow users to upload their own papers and extract insights.
• Add Long Term Storage: Allow users to upload their own papers and extract insights.
• Citation Generator: Export results in APA/MLA/BibTeX format.
• LLM Model Swapping: Use Anthropic Claude, Mistral, or open models like LLaMA via HuggingFace. You could even train models for each specific module.
• Multi-language Support: Let users input questions in other languages.
• User Accounts: Save searches and generated reports to a personal dashboard.

License

This project is licensed under the MIT License. See the LICENSE file for details.