Editorial policy
This page sets out how ShilajitDB decides what to publish, where its health claims come from, and what happens when it gets something wrong. Last reviewed 27 October 2026.
Who is responsible
ShilajitDB is an independent publication. Its editors do not publish their names. What you can check instead is the fixed grading method, the lab report behind each grade (linked from the product page), the paper behind each health claim (listed on /research) and a dated corrections and changes log. Corrections, data and press: hello@shilajitdb.com.
Nothing on this site is medical advice. It does not recommend shilajit for any condition or a dose for treating one.
What we publish
- Product grades. A transparency grade, a quality tier and an overall grade for each product, from the documents the brand publishes.
- Ranking pages (/best). Generated by rule from those grades. Where a ranking involves an editorial choice, the page says so.
- Research guides (/learn). Plain-language articles on what the studies show and what a lab report can and cannot tell you.
How products are graded
Grades come from a fixed formula that scores what a lab report documents: who issued it, whether it gives measured heavy-metal values on the finished product, whether it names the laboratory, includes a microbial panel and a batch code, and how recent it is. We read each report ourselves rather than relying on what a brand says about it. Claims with no document behind them (“third-party tested”, “GMP certified”) are recorded as claims and do not score.
The formula is the same for every brand. No brand gets a different formula, a manual override or paid placement. The grades measure documentation, not efficacy, and a low grade means missing evidence, not proof of contamination. Full details: methodology.
How we source health claims
- Every paper has an entry. Each study we cite is listed on /research with a finding and a caveat written from the paper’s own abstract. If an article makes a claim the abstract does not support, the entry is marked unverified and the claim is left out or flagged.
- Identifiers are checked. Articles cite through a registry, so a citation to an unknown paper fails the build, and a script compares each paper’s author and year with PubMed.
- We say what kind of evidence it is. A human trial, an animal study and a test-tube result are labelled as different things, and where no human trial exists we say so plainly.
- When a citation changes, we read the paper. Fixing an identifier without reading the paper is how errors are introduced; see how we audited our citations.
Independence and money
ShilajitDB earns no affiliate revenue, and no brand pays for grades, reviews or placement. The outbound links to brands and retailers are plain links. See the disclosure.
Brands and requests
A brand can send us a lab report. It is read against the same rubric as every other, and if it changes the grade the change goes in the log. We correct errors of fact. We do not remove or soften an accurate grade on request, and we do not offer a better grade, or coverage, in exchange for anything.
When we are wrong
We log corrections with the date, what was wrong and what changed, including errors that made a product look better than it should have. The log also lists the problems we know about and have not yet resolved.
What the database does not cover
Products are catalogued from brand websites, Amazon and UK retailers, not every shilajit product sold anywhere. A brand may hold a report we could not find. Articles carry a “last reviewed” date, and ranking pages and the data downloads are rebuilt from the database hourly.