Integrative Oncology Research

Your doctor has 12 minutes. The internet has an agenda. I have time, and I work for you.

You're navigating one of the hardest things a person faces. You deserve real answers, not noise.

VERA is an AI research analyst that evaluates complementary cancer therapies using a rigorous evidence hierarchy — so you can walk into your next appointment informed.

See the demo → Watch the explainer
What VERA does.

VERA doesn't work for a hospital, a supplement company, or anyone with a stake in what you decide. She works for you — and her only job is to help you ask the right questions.

See the demo
Watch a real VERA session, end to end

A scripted recreation of an actual session — a Stage IV pancreatic patient asks about high-dose IV vitamin C during FOLFIRINOX. Watch the scoping interview, the live source-by-source research, and the final report card. No buttons to press — it runs itself.

Open the 6-page PDF → This is the actual report the chat above generates — the deliverable that goes to the patient's oncology appointment.
How it works
The methodology, named

VERA isn't a chatbot wrapping a search engine. She runs a five-stage research protocol — every stage with named, file-enforced rules and an artifact written to disk. Skip any stage and the next one refuses to start.

1
Scoping — she refuses to answer until she knows enough to answer well

One question per turn, VERA establishes the four scoping fields: cancer type and stage, current treatment, what you're trying to evaluate, and your contraindication context (medications, allergies, biomarkers, weight). Urgency does not skip scoping. "Just tell me — does it work?" gets the same calm "I need one more thing first." The scoping summary is the contract for everything downstream.

2
Retrieval — six academic databases plus the same web your patient is reading

A 6-source fallback chain runs in order — Europe PMC, PubMed, ClinicalTrials.gov, Semantic Scholar, Unpaywall, WebSearch — and a parallel community scan covers Reddit, patient forums, and podcast transcripts so VERA knows what claims you're hearing online, not just what's published. Both sides of the literature are in the evidence ledger before any synthesis begins.

3
Weighting — every source gets a tier, T1 RCT to T6 anecdote

A 6-tier hierarchy tags every source before it can be cited: T1 systematic reviews and meta-analyses, T2 randomized controlled trials, T3 observational studies, T4 case reports, T5 grey literature, T6 anecdotal. A source without a tier doesn't enter the ledger. Tier drives weight; weight drives the verdict.

4
COI flags and adversarial audit — who paid, who profits, who's selling

Every Tier 1 and 2 source is checked for industry funding, author conflicts, and structural bias (geographic capture, ideological capture, publication bias). Then a second AI sub-agent runs the audit-brief protocol against the finished ledger — looking for missed searches, weak sources cited as strong, hand-waved gaps. Phase 3 cannot start until the audit clears.

5
Deliverable — two audiences in one PDF

A 6-page report. Page 1 hits the patient first: verdict in two sentences, a two-question dashboard ("will it help / could it hurt"), and the counts that drove the answer. Pages 2–4 are patient-facing at sixth-grade reading level — findings, gaps, and the specific questions to ask your oncologist. Pages 5–6 are the technical brief for the doctor: tier-tagged evidence, COI flags, hyperlinked references. Print it. Walk in informed.

What VERA does and doesn't do
Built around your interests — not the institution's
VERA researches evidence

She evaluates the published literature on complementary therapies with consistent, rigorous criteria — the same standard for everything she reviews.

VERA does not give medical advice

VERA is a research tool, not a clinician. She surfaces evidence; she does not interpret it for your specific medical situation.

VERA does not make treatment recommendations

No "you should take this." VERA presents what the evidence says and where it's strong, weak, or absent — full stop.

VERA works for the patient

No pharma relationships. No institutional bias. No agenda. VERA's only objective is to help you ask better questions.

For judges and developers

Clone, run, evaluate

View on GitHub Judge guide