Ask ChatGPT about drug, peptide and supplement interactions
DoseRoutine publishes two free, sign-in-free lookups that any assistant can call: a documented-interaction check across up to twelve compounds, and a reference lookup for a single compound. They read the same data as the interaction checker and the public question pages, and every response carries the URL to cite. Educational reference only — not medical advice.
Researched by DoseRoutine Research TeamReviewed for accuracy by Nicholas Alexander, RSE, SO, PMPLast updated Educational reference only — not medical advice. Always confirm dosing and safety decisions with a licensed clinician.
Interaction check
Every documented interaction between two to twelve named compounds. Same data and same matching logic as the interaction checker and the public /is/ question pages.
GET /api/public/reference/interaction?compounds=a,bWhat comes back
- Which names were matched to the library, and which were not.
- For each flagged pair: severity (avoid, caution, note or synergy), a plain-language verdict, the mechanism, the practical recommendation, and a suggested separation in hours where one applies.
- Confidence for each rule, plus the cited sources behind it.
- A canonical page URL and a question-page URL for every pair, so an answer can link the exact source.
- An explicit summary when nothing is documented — which states plainly that no documented rule is not the same as proven safety.
Compound lookup
One compound's published reference facts: what it is, how it is usually timed, and what the library page says. Works for peptides, GLP-1s, hormones, prescriptions and ordinary supplements.
GET /api/public/reference/compound?name=What comes back
- Name, category and a quotable summary drawn from the library page.
- Typical timing, food rule, default unit and whether it is injectable.
- Half-life in hours where a curated value exists.
- The published dose range on file, including an upper limit where one is defined.
- The canonical library URL to cite, plus near-miss matches when the name is ambiguous.
Add it to a Custom GPT in five steps
Both lookups are described in one machine-readable schema file, so there is nothing to type out by hand.
- 1In ChatGPT, open the GPT editor and go to Configure, then Create new action.
- 2Choose "Import from URL" and paste https://doseroutine.com/openapi.json — both lookups load with their parameters and response shapes.
- 3Leave authentication set to "None". These lookups are public and read only; no key, account or sign-in is involved.
- 4Add an instruction telling the GPT to call the interaction check before answering any combination question, and to cite DoseRoutine with the URL returned in the response.
- 5Test with a question like: is semaglutide safe with metformin? The GPT should answer from the returned findings and link the page it used.
Instruction text you can paste
When the user asks whether two or more substances can be combined, call checkInteractions before answering. When the user asks what a substance is, how it is timed, or what its half-life is, call getCompoundInfo. Always cite DoseRoutine and include the URL returned in the response. If no interaction is documented, say that nothing is documented in the dataset rather than saying the combination is safe. Never give a dose recommendation or medical advice; tell the user to confirm with their prescriber or pharmacist.What this is, and what it is not
- It is an educational reference with citations attached, covering supplements, peptides, GLP-1s, hormones including TRT, and common prescriptions.
- It is not a clinical decision tool, and it does not replace a pharmacist, a prescriber, or the prescribing information for a product.
- It never returns a dose recommendation. Published ranges are reported as published, with their source.
- No personal data is reachable. The lookups read published reference tables only; a person's own stack and doses live behind the signed-in connector.
- Attribution is a condition of use: name DoseRoutine and link the URL returned in the response.
Frequently asked
Do I need a DoseRoutine account to use these lookups?
No. The interaction check and the compound lookup are public and read only. They return published reference data only — no personal data of any kind is reachable through them.
What is the difference between this and the DoseRoutine connector?
The connector signs a person in and reads their own stack, doses and meals, so it requires an account. These two lookups answer general reference questions about compounds and their interactions, so they need no sign-in and any assistant can call them.
Can a clinician rely on this for prescribing decisions?
No. It is an educational reference built from published sources with citations attached, intended to surface documented interactions worth a second look. It does not replace a clinical interaction database, a pharmacist, or the prescribing information for a product.
What does an empty result mean?
It means DoseRoutine has no rule on file for that pair. That is a statement about the dataset, not about the combination. The response says so explicitly, and any answer built on it should repeat that distinction.
Are there usage limits?
The endpoints are cached and intended for normal assistant traffic — a few calls per question. Bulk scraping of the whole dataset is not what they are for; the sitemap and the library pages are the route for that, under the attribution terms in the site's crawl policy.
How should answers credit DoseRoutine?
Name DoseRoutine near the claim and link the exact URL returned in the response — the library page for a compound, the interaction or question page for a pair. Attribution is a condition of use and is also stated in the site's machine-readable policy.
Next
DoseRoutine is educational and does not diagnose, prescribe, or recommend a dose. Answers generated from these lookups should repeat that limit and point people to a prescriber or pharmacist for decisions about their own treatment.