Our Methodology
Our Goal
Our mission is to make research more accessible by simplifying access to scientific information and demonstrating that AI can bridge the gap between complex topics and a wider audience.
Agentic AI for Reports Generation
To streamline the process of gathering and synthesising research data, we use an agent-driven pipeline:
- A large language model is prompted to retrieve comprehensive information on research efforts aimed at curing rare diseases.
- The agent has access to a web search tool and can parse the results to extract relevant insights.
- The website is open source; the generator pipeline can target any OpenAI-compatible model, including a local one.
Open-Source
This project is entirely open source. UpToCure GitHub Repository.
Built thanks to numerous open-source projects, with a special emphasis on the Open Deep Research framework.
Limitations
- Risk of hallucinations: LLMs can occasionally produce inaccurate or fabricated information. Always verify with the cited sources.
- Recent research gaps: very recent scientific developments may not yet be referenced on the web.
- Translation accuracy: reports are initially generated in English and then translated to other languages.