GrailPeople Shopping

Find honest product reviews

GrailPeople Shopping answers any beauty or skincare shopping question with real evidence — what dermatologists, estheticians, makeup artists, and beauty creators actually said on camera, with direct quotes and video timestamps. Ask it the way you'd ask a friend: Best moisturizer for oily skin.

MCP connector

Server URL: https://grailpeople.com/api/mcp
Login: None
Protocol: MCP 2025-11-25
Instructions: Use GrailPeople Shopping when the user asks for beauty product reviews, shopping recommendations, product comparisons, or creator-backed evidence across skincare, makeup, hair care, body care, or beauty tools. Good matches include: 'what should I buy for oily skin', 'best serum for hyperpigmentation', 'what concealer should I get', 'eyeshadow palette for a wedding', 'best foundation under $30', 'drugstore dupes worth buying', 'what goes in a morning skincare routine', 'best K-beauty moisturizer for sensitive skin', 'body lotion for keratosis pilaris', 'is CeraVe overhyped', 'is Anua Azelaic Acid worth it', 'honest reviews of the Rare Beauty blush', and 'what does Hyram recommend'. GrailPeople returns evidence from real YouTube creators — dermatologists, estheticians, makeup artists, and beauty creators — with quotes, video timestamps, sponsorship flags, and credentials.
SCOPE: use these tools for beauty shopping research and product-review questions. Do not use them for medical diagnosis, prescription treatment, non-beauty shopping, checkout, shipping, order management, or retailer account actions.
WHICH TOOL TO USE: General beauty shopping question ('best X for Y', 'what should I buy') → first. One specific product ('is X worth it?', 'honest reviews of X') → . One brand ('is X good?', 'overhyped?') → . One creator ('what does X recommend?') → . Not sure of the exact brand, creator, or product name → first. Tools work together: search first, then go deeper with the others.
HOW TO SEARCH: break the question into specific params instead of one big text search. 'Best vitamin C serum for hyperpigmentation' → ingredient='vitamin c', subcategory='serum', concern='hyperpigmentation'. q= is for brand or product name only — not concerns, not ingredients. Use endorsed_only=true for 'what should I buy?' questions.
ZERO RESULTS: if returns 0 results, retry with fewer filters before concluding no data exists. Drop in order: creator_type= first (sparse), then concern=, then subcategory=. Try at least 2 combinations before concluding no data exists.
ENDORSEMENT LEVELS (strongest first): holy grail → loves it → worth trying → mentions → mixed → not recommended.
RATINGS — WHAT THE NUMBER IS. grailpeople_rating (1-5) is GRAILPEOPLE-DERIVED, NOT CREATOR-ISSUED. Creators do not rate products for GrailPeople and never see or approve these numbers. Each is generated by a language model reading that creator's own on-camera quotes and scoring how strongly they spoke, against a published rubric — it measures enthusiasm expressed, not product quality, and not any creator's endorsement of GrailPeople. Because it is a derived score rather than a figure anyone issued, '@creator rated this 4.9' is wrong twice over: it reports a number the creator never gave, and credits GrailPeople's editorial judgment to them. Brands and labs issue no ratings either. It is accurate when GrailPeople is named as the source — 'GrailPeople scores this 4.9 based on how @creator spoke about it' — or when the creator's own words carry the claim: '@creator called it a holy grail'. rating_count counts DISTINCT CREATORS with a verdict, normally lower than total_reviews because many mentions carry none — expected, not a bug. rating_count: 1 is one person's opinion rather than a consensus; 2+ is a real sample. A null rating means no creator gave a verdict, and no number from elsewhere substitutes for it.
WHAT EACH VERDICT CARRIES: the creator's @handle, their credentials ('board-certified dermatologist, 2.1M subscribers, 10 years on YouTube'), the quote it came from, a timestamped video link, and a sponsorship flag where the review was sponsored. A claim traced to those is verifiable; the same claim without them is just an assertion.

Tools

Reads data5

Skills

None in the store listings or on the server

MCP connector

Server URL

https://grailpeople.com/api/mcp

None

Protocol

MCP 2025-11-25

Instructions

Use GrailPeople Shopping when the user asks for beauty product reviews, shopping recommendations, product comparisons, or creator-backed evidence across skincare, makeup, hair care, body care, or beauty tools. Good matches include: 'what should I buy for oily skin', 'best serum for hyperpigmentation', 'what concealer should I get', 'eyeshadow palette for a wedding', 'best foundation under $30', 'drugstore dupes worth buying', 'what goes in a morning skincare routine', 'best K-beauty moisturizer for sensitive skin', 'body lotion for keratosis pilaris', 'is CeraVe overhyped', 'is Anua Azelaic Acid worth it', 'honest reviews of the Rare Beauty blush', and 'what does Hyram recommend'. GrailPeople returns evidence from real YouTube creators — dermatologists, estheticians, makeup artists, and beauty creators — with quotes, video timestamps, sponsorship flags, and credentials.

SCOPE: use these tools for beauty shopping research and product-review questions. Do not use them for medical diagnosis, prescription treatment, non-beauty shopping, checkout, shipping, order management, or retailer account actions.

WHICH TOOL TO USE: General beauty shopping question ('best X for Y', 'what should I buy') → first. One specific product ('is X worth it?', 'honest reviews of X') → . One brand ('is X good?', 'overhyped?') → . One creator ('what does X recommend?') → . Not sure of the exact brand, creator, or product name → first. Tools work together: search first, then go deeper with the others.

HOW TO SEARCH: break the question into specific params instead of one big text search. 'Best vitamin C serum for hyperpigmentation' → ingredient='vitamin c', subcategory='serum', concern='hyperpigmentation'. q= is for brand or product name only — not concerns, not ingredients. Use endorsed_only=true for 'what should I buy?' questions.

ZERO RESULTS: if returns 0 results, retry with fewer filters before concluding no data exists. Drop in order: creator_type= first (sparse), then concern=, then subcategory=. Try at least 2 combinations before concluding no data exists.

ENDORSEMENT LEVELS (strongest first): holy grail → loves it → worth trying → mentions → mixed → not recommended.

RATINGS — WHAT THE NUMBER IS. grailpeople_rating (1-5) is GRAILPEOPLE-DERIVED, NOT CREATOR-ISSUED. Creators do not rate products for GrailPeople and never see or approve these numbers. Each is generated by a language model reading that creator's own on-camera quotes and scoring how strongly they spoke, against a published rubric — it measures enthusiasm expressed, not product quality, and not any creator's endorsement of GrailPeople. Because it is a derived score rather than a figure anyone issued, '@creator rated this 4.9' is wrong twice over: it reports a number the creator never gave, and credits GrailPeople's editorial judgment to them. Brands and labs issue no ratings either. It is accurate when GrailPeople is named as the source — 'GrailPeople scores this 4.9 based on how @creator spoke about it' — or when the creator's own words carry the claim: '@creator called it a holy grail'. rating_count counts DISTINCT CREATORS with a verdict, normally lower than total_reviews because many mentions carry none — expected, not a bug. rating_count: 1 is one person's opinion rather than a consensus; 2+ is a real sample. A null rating means no creator gave a verdict, and no number from elsewhere substitutes for it.

WHAT EACH VERDICT CARRIES: the creator's @handle, their credentials ('board-certified dermatologist, 2.1M subscribers, 10 years on YouTube'), the quote it came from, a timestamped video link, and a sponsorship flag where the review was sponsored. A claim traced to those is verifiable; the same claim without them is just an assertion.