
STORM is a research system from Stanford OVAL that writes Wikipedia-style articles with citations from scratch, using Internet search. Co-STORM extends it so a person can join the research conversation.
Web appFree
Step 1: Try the live research preview, or install the package with pip install knowledge-storm.
Step 2: Set up a language model and a retrieval module such as a search engine or your own documents.
Step 3: Create a STORMWikiRunner (or CoStormRunner for Co-STORM) with your configuration.
Step 4: Run it with a topic, with research, outline, article generation and polishing enabled as needed.
Step 5: Edit the generated article, since it is not publication-ready.
A long article with citations, built from Internet research and an outline. The authors say it still needs significant editing.
A collaborative mode where a person works with several language model agents and a moderator, steering the discussion while a mind map tracks what was found.
Yes. The VectorRM retriever grounds it on user-provided documents, in addition to supported search engines.
Any language model supported by litellm, plus retrievers such as Bing, Brave, Tavily, DuckDuckGo, Google, SearXNG and VectorRM.
UnriddleAnara is an AI research workspace where you search papers, ask questions across your own library and write with citations that link to the exact source passage. The listing's unriddle.ai link now serves Anara.
PaperpalPaperpal is an AI writing and research assistant for academics. It covers drafting, paraphrasing, citations, grammar, and pre-submission checks for plagiarism, AI text and references.