AI Blocks › Research a topic properly
Scientific web searcher
Name a research topic. Get papers and pages found by meaning, with text pulled out of the links. Uses Exa.
Scientific web searcher is a free AI skill for the job "Research a topic properly". Give the instructions to an AI assistant that supports the required files and tools, then review its output.
Free · No account · MIT licence · Made by K-Dense · Thinking: low
Made by K-Dense. See the original on GitHub; we copied it as it is, with its licence.
What it does
A set of papers and web pages on your topic, found by meaning instead of exact keywords, with the page text already pulled out so Claude can read it. It runs on Exa, a search service, so you need an Exa API key and Python. For science and technical questions it runs two passes. The first uses a scholarly bias to favour journals, preprints and trusted institutions. The second looks at a short list of academic sites. It prefers peer-reviewed work, then preprints, then government and university pages, over blogs. When you already have links, it can fetch the full text of many pages, including PDFs, in one batch.
You give it
- A research topic or question
- Links to pages or PDFs to read (optional)
- An Exa API key
- Domains to prefer or avoid (optional)
You get
- A ranked list of papers and pages
- Extracted text from each link
- Sources sorted by how scholarly they are
- Full text of batches of URLs
Use it when
- You are starting a literature search and want relevant papers fast.
- You have a list of article links and need the text from all of them at once.
- You want current findings on a technical topic without wading through blog posts.
- You need to find preprints when no peer-reviewed version exists yet.
How to use it
- Add it to your AI once: download it and add the folder to Claude, ChatGPT or Codex as a skill.
- Ask for the job in your own words. Your AI opens the skill when it fits.
- It works best when your AI can run code and call the Exa search service.