Petal vs Scholarly

Petal is a web-based research assistant and reference manager for academic work, also available as a browser extension. Its plans include a free tier and paid options listed from $4.99/user/mo (annual). It helps users narrow literature reviews and compare studies, while providing a library for uploading and organizing PDFs with collections, metadata, and tags. AI tools offer translation, summaries, explanations, key-point extraction, suggestions for questions and search terms, and conversational assistance. Shared workspaces support group projects, co-authored papers, and reviews. The citation generator supports more than 10,000 styles and can copy citations or export them to BibTeX or Word. References can be brought in through identifiers such as DOI, PMID, ArXivID, ISBN, ISSN, and EAN, or by capturing metadata from URLs. Browser add-ins are listed for Chrome, Firefox, and Safari, alongside a Word add-in for citations and bibliographies. The free plan includes 1 GB storage, one seat, three guests, 400 AI credits, two collections, and three annotations per document; citation export is not included.

Scholarly is a browser-based AI workspace that uses your source material to create study and research outputs. Add PDFs, recordings, notes, slides, webpages, videos, images, YouTube links, or pasted text, then make flashcards, quizzes, practice exams, study guides, podcasts, video lectures, slides, mind maps, worksheets, and timelines. Chat answers include checked citations, with a Sources button opening the page supporting each claim. Deep Research searches the web, evaluates sources, and produces cited reports with charts and inline links.

Scholarly works in a browser on laptops, tablets, and phones, and syncs work through your account. Its Google Drive connection can use selected files as sources and save finished work as new Docs or Slides. Teams can use shared libraries, roles, central billing, and model controls. A free plan and free trial are available; paid plans start at $12/user/mo (annual). Scholarly says private user content is not used to train general-purpose AI by default; training requires separate opt-in.