chatgenius_stats
ChatGPTCount what visitors asked ChatGenius, the AI assistant on the customer's website: answered questions, conversations and distinct visitors per day, week or month, with totals (including liked and disliked answers) and an optional breakdown by country, state or page. Use it for any how-many or trend question, e.g. "questions in the past 30 days, day by day", rather than paging through questions. Days are counted in the account's timezone, named in the result.
check_leads_credits
ChatGPTHow many leads credits this account has used and has left in the current subscription period, and when the period ends. One credit is one contact in a finished export; previews are free. Use it whenever the user asks what they have used or have left, or before a large export. Checking is free. Report credits_remaining as given; if exports_in_progress is present, say the balance will drop once those finish.
citation_stats
ChatGPTCitation counts for the account: all-time, this calendar year, and this month. Pass any search filter to also count only matching papers - this is how to answer questions like "how many citations did product X get in 2025?" without paging through results. Note total_all_time counts searchable papers and will match search_citations; dashboard_total is the larger figure shown on the customer's own dashboard, which also counts approved papers that lack a title or publication year. If both appear, they are both correct.
citation_widget_searches
ChatGPTWhat visitors typed into the citation widget's search box on the customer's website: each term with how many distinct visitors searched it, how many times, and when it was last searched, most visitors first. A direct demand signal for what people look for on the customer's site.
citation_widget_stats
ChatGPTCount how visitors used the citation widget embedded on the customer's own website (the "cited by N papers" badge): distinct visitors, clicks on cited papers, and searches, per day, week or month, with totals and an optional breakdown by visitor country or state. Use it for any how-many or trend question about the widget, e.g. "paper clicks in the past 30 days, day by day". Days are in the account's timezone, named in the result. This is the embedded widget's traffic, not the citation library itself (for the library use citation_stats / search_citations).
citation_widget_top_papers
ChatGPTThe cited papers website visitors clicked most in the citation widget, highest first, each with its title, journal, year, the customer's product it cites, and the pages on the customer's site where it was clicked. Use it for "which papers are visitors most interested in" / "what drives engagement on our site". paper_id works with get_citation.
deep_analysis
ChatGPTStart a Deep Analysis of a research field and return an analysis_id. Deep Analysis pulls every PubMed paper matching a keyword and computes the landscape of the field: publication trend by year and a growth score, the leading journals (with impact factors), authors, institutions, hospitals, companies and countries, the top diseases / genes / brain regions, key phrases, and the top papers. Use it when the user wants the SHAPE of a field -- trends, who leads it, which journals matter, background for a review -- rather than a list of individual papers (for those, use the search tools). The analysis runs in the background (about 15-60 seconds); then call get_deep_analysis with the analysis_id to read the result.
export_citations
ChatGPTExport citations as a file. Use this whenever the user wants "all" of something, a spreadsheet, a bibliography, or more results than are worth reading in conversation - paging through hundreds of results with search_citations wastes the conversation and will not finish. By default it produces the customer's own BizGenius spreadsheet - the same columns as their Library page export - built in the background and EMAILED to the account holder as a download link, usually within minutes; it does not come back into the conversation. format=bibtex instead returns a BibTeX download link right away, for reference managers. Takes the same filters as search_citations; omit them all to export the whole library.
export_leads
ChatGPTQueue the full contact list for a query - up to 5,000 contacts - as a spreadsheet emailed to the account holder, usually within a few hours. Each contact in the finished list uses one leads credit (check_leads_credits shows the balance), so only call it when the user asks for a list or an export, and preview with search_leads first. Takes the same arguments as search_leads. The list does not come back through this conversation.
get_chatgenius_conversation
ChatGPTOne visitor conversation with ChatGenius in full: every question and complete answer in order, with times, pages and ratings, and the citation-library papers the assistant was given to answer from. Get the conversation_id from the other ChatGenius tools.
get_citation
ChatGPTFetch a single paper in full by its BizGenius id, including the complete abstract, the AI summary, tags, and volume/issue/pages. Get the id from search_citations.
get_deep_analysis
ChatGPTRetrieve a Deep Analysis by its analysis_id (from deep_analysis or list_deep_analyses). While it is still computing, the result has status "running" -- poll again every ~15-20 seconds until status is "done". A finished result holds the publication trend and growth score, and the top journals, authors, institutions, hospitals, companies, countries, diseases, genes and key phrases of the field. The top papers are included when the result is fetched shortly after it finishes.
list_chatgenius_conversations
ChatGPTVisitor conversations with ChatGenius, most recent first: when, how many questions, the first question, the page, country and IP address, any email address the visitor left, ratings, and whether the IP is blocked. Use it for who-and-where questions such as "which visitors left an email this week". Filters match a conversation when any of its questions matches.
list_deep_analyses
ChatGPTList this account's past Deep Analyses, newest first, each with its analysis_id, keyword and date. Use get_deep_analysis to open one.
list_leads_exports
ChatGPTThe account's recent leads exports - queued, running or ready, with a download link once ready. Use it for "is my list ready" or "send me the link again". For the credit balance use check_leads_credits.
list_new_citations
ChatGPTPapers most recently added to the library, newest first. Use this for "what is new this week/month" questions. Note this orders by when BizGenius found the paper, which is not the same as when it was published.
search_chatgenius_questions
ChatGPTThe questions website visitors asked ChatGenius, newest first, each with the start of the answer it gave, the page it was asked on, the visitor's country, IP address, any email address typed into the question, and the visitor's rating. Use it for what-are-visitors-asking questions: topics, product interest, questions answered badly. For counts use chatgenius_stats; for a whole conversation, get_chatgenius_conversation.
search_citations
ChatGPTSearch the papers that cite this customer's products. Returns a compact record per paper including the snippet - the sentence from the paper where the product is actually mentioned, which is usually the most useful field for answering questions. Abstracts are truncated here; use get_citation for the full text.
search_leads
ChatGPTFind researchers to contact about a product: a PREVIEW of the contacts (name, email, affiliation, and the paper or grant they were found on) on the first matching papers or grants. Use it to check a query finds the right people before exporting. It is not the full list - export_leads is.
top_citing_organizations
ChatGPTThe universities, companies, or journals recorded as citing this customer. Answers "who is citing us". Returns organization names, plus a logo URL where one is available. Note BizGenius does not track a per-organization citation count, so this is a set rather than a ranked list.