compare_blood_tests
ChatGPTUse this when the user explicitly asks to compare two to four known Praetor Health blood tests. Supply each canonical slug or exact slugified published title. Call it at most once. Do not call search_blood_tests or get_blood_test in the same turn, and do not call this after a discovery result because search_blood_tests already returns the public facts needed for an inline comparison. If the products cannot be identified by slug or published title, search once and answer from that result instead. Compare public product facts without declaring a medically best test. It does not diagnose, determine suitability, order a test or book a clinic.
compare_tests
ChatGPTUse this to compare two to four Praetor tests: markers they share, markers only in each test, whether one contains another, and the price and marker-count differences. It returns data only; call show_tests with view compare to display the comparison. It compares facts and never says which test is medically best.
get_biomarker
ChatGPTUse this when the user asks about one biomarker in connection with Praetor's tests: it returns Praetor's public summary, the marker's category and related markers, a link to the praetor.health guide, and the tests that include it. It returns data only: always follow it with show_tests (view biomarker, the same biomarker) so the person sees the card and the tests that include it. It is educational information, never an interpretation of anyone's results.
get_blood_test
ChatGPTUse this when the user explicitly asks to open or show details for one known Praetor Health blood test. Supply its canonical slug or exact slugified published title. Call it at most once. Do not call search_blood_tests or compare_blood_tests in the same turn, and do not use this after a discovery result merely to support an answer. It returns public catalogue facts only; it does not access customer accounts, orders, results or reports or determine medical suitability.
get_test
ChatGPTUse this for the full public facts of one Praetor test: markers by category, preparation, every collection option with its total, turnaround and the neighbouring tiers. It returns data only; call show_tests with view detail to display the test. Accepts the slug or the published name written as a slug.
list_tests
ChatGPTUse this to read Praetor Health's whole public UK catalogue in one call: every test, grouped by range, with who it is for, its test price, the lowest total including collection, collection methods, turnaround and markers. Choose the tests whose markers fit what the person wants to check, for any goal, topic, budget or collection preference, then call show_tests with their slugs, or with view families and range ids for a broad overview. Pass a budget as maxPriceMinor and a required collection method as collectionMethods, however the person words them; eligibility leaves out the other sex's tests. Show tests from a filtered list with view results, not families, which shows whole ranges. Reuse a full list for follow-up questions; call it again when the person's limits change. It draws nothing. These are product facts, not diagnosis, personal suitability or medical advice.
load_view
ChatGPTWidget-only: loads the data for another view inside the Praetor explorer.
search_blood_tests
ChatGPTUse this when the user asks to browse or find public Praetor blood tests by product name, biomarker, published category, collection, eligibility, budget, goal or health topic. The tool returns product facts only; ChatGPT decides whether a catalogue lookup is relevant and handles the conversation. An app mention or greeting alone needs clarification, not a search. Set query to concise catalogue search terms, not the full user message; omit personal identifiers, test results, medication history and urgent-care details. Use mode browse for a catalogue-wide listing; otherwise use search. Restrict categories only when the user asks for a published product family; do not invent categories for biomarkers or conditions. Optional biomarkers select products containing every requested marker. Preserve explicit budget, collection and eligibility constraints. Results contain the public facts needed for an inline comparison. Normally make one search and do not follow it with get_blood_test or compare_blood_tests in the same turn. On an input error or no_matches, one corrected search is allowed using the returned schema or catalogue categories; never silently remove a user constraint. An empty search does not mean Praetor has no tests. For an explicit comparison of two to four known products, call compare_blood_tests directly with canonical slugs or exact slugified published titles. For one known product detail request, call get_blood_test directly. These tools provide catalogue facts, not diagnosis, interpretation of personal results, treatment guidance or individual medical suitability. They do not order a test or book a clinic.
show_tests
ChatGPTUse this to display Praetor tests after a data tool: a families overview, the tests you chose (one range shows as a tier ladder), one test, a comparison of two to four tests, or one biomarker. Pass identifiers only (slugs, range ids, a biomarker name); every fact shown is read from Praetor's catalogue. Call it once per answer. The view lets the user compare, see what one test adds, open markers and continue to praetor.health to buy.
