Useful
Research
Brief a paper, check what a source really supports, and pull claims out with the quote attached.
How Research agents work
You put in
Brief a paper, check what a source really supports, and pull claims out with the quote attached.
Agent
Research
You get out
A prompt you can copy, a file you can download, and the steps to run it today.
A tab of papers and posts is not a brief. The useful part is small: what the source claims, which sentence supports it, and what you are not allowed to say yet because you did not read that section. This shelf is for people doing that reading with a model in the other window, who still want the citation to survive a second look.
Paper to Brief turns an abstract and your notes into a page you can hand someone, and it marks what you did not read. The credibility check takes a claim you hope is true and the source text, then says what the source actually supports. The claim extractor pulls claims out of an article or a transcript and attaches a quote, plus a question you still need to check before you repeat it.
Quotes have to come from the paste. If the model cannot point at a sentence, it does not get to keep the claim. These agents do not browse the web, do not invent a DOI, and do not upgrade a blog post into a study. They are a reading desk, not a search engine.
Questions about Research
Will the agent fetch the paper for me?
No. Paste the abstract, the section, or the notes you took. If a passage is missing, the brief should say it is missing rather than fill the gap from memory.
Can I quote the output in a published piece?
Quote the source, not the agent's paraphrase, and check the quote against the original. The prompt is there to keep the paraphrase honest, not to be the citation.
What if the source is a tweet or a marketing page?
Paste it anyway and say what it is. The credibility check is stricter on a claim that a thin source cannot carry. It will not promote the source to a study.