Which Ayurvedic manufacturers publish what is in the bottle

Disclosure, before anything else. This site is published by Age Ayurveda, which sells one of the products counted on this page. That is a conflict of interest and you should read this page knowing it. What we have done about it: the comparison is organised by preparation rather than by brand, every count below is of what the company itself published rather than our judgement of it, we are in the tables and we come last, and we have not scored, ranked or picked a winner. If you find an error about any company here, including a competitor, tell us and we will correct it.

A classical Ayurvedic preparation is not a brand’s recipe. It is a formula from a classical text, and in India the Ayurvedic Formulary of India states each one officially: the ingredients, their plant parts, and the quantity of each. The companion page on what the formulary fixes sets out exactly what that binds.

So there is a published standard, and a reader holding a bottle has a fair question: does what is in this bottle match it? That question is answerable only if the manufacturer publishes its side. This page measures whether they do.

What was measured, and what it means

Every company’s robots.txt was fetched and parsed first. Where it declined this crawler, nothing was fetched and the refusal was recorded. Where it allowed, the catalogue was enumerated from the company’s own sitemap and each product page was read once, with a gap of at least a second and a half between requests to any one host and any stated Crawl-delay honoured.

Two populations, counted separately, because they answer different questions. Both were read on 9 October 2026. The first is every product on each permitted host, classical and proprietary alike: 6,086 pages read of 6,088 attempted, across 16 of the 20 companies surveyed. That is the figure that is a fact about a company. The second is the subset sold under one of the 101 formulary names transcribed on this site, 474 pages across 14 companies. That is the figure that bears on whether a product can be checked against the entry it claims, which is what this page is about. For most companies the two agree closely. For two of them they do not, and the tables below show both rather than picking one.

The four companies with no measured range, and the reason for each, are listed under the tables. Two of those four are findings about them; one is a finding about the limits of our own matching; one is a crawler refusal we obeyed.

Each page falls into one of four states, and the distinction between the second and the third is the whole point:

• Ingredients and quantities. The page lists what is in the preparation and how much of each. This is the only state in which a reader can check the product against the formulary entry.
• An ingredient list only. The page names the ingredients and not their amounts. You can see whether an ingredient is present or absent; you cannot see whether the proportions are the formulary’s.
• Neither. The page carried substantial readable text and no ingredient list in it.
• Could not be read. The page served almost no text, which means its content renders in the browser rather than in the HTML. That is a fact about a company’s stack and not about its disclosure, so it is never merged into the one above. 11 pages on two hosts are in this state.

Two pages of the 6,088 could not be fetched at all. They are excluded from every count rather than counted as publishing nothing.

Everything each company sells

The whole product range on each host that permitted it, classical and proprietary alike. This is the figure that is a fact about the company.

Only the products sold under a formulary name

The same companies, counting only products whose name is one of the 101 formulary entries transcribed on this site. This is the figure that bears on whether a product can be checked against the entry it claims, and for two companies it is markedly higher than the first table.

Four further companies were surveyed and contributed no measured page. Each reason is itself a finding.

• Arya Vaidya Sala Kottakkal: its robots.txt declines this crawler, and declines every AI crawler tested except Googlebot, so nothing was fetched.
• Nagarjuna Herbal Concentrates: its sitemap is readable and lists no individual product page at all, only category listings, so there is no product page of theirs for a composition to appear on.
• Unjha Pharmacy: no sitemap was found, and all 2 catalogue links on its own home page return 404, so its catalogue cannot be reached from its own site.
• Vaidyaratnam Oushadhasala: its catalogue lists its products as top-level page names that cannot be told apart from its ordinary site pages, so its individual product pages could not be identified here. This is a limit of our own matching, not a statement about what it publishes.

What each company published, by preparation

Only the preparations sold under a formulary name by three or more of the companies read, which is 63 of them. The second column is the formulary entry itself, as transcribed and published on this site, which is the thing any of these products can be checked against.

Every cell was read on 2026-10-09. “None” means no company read here was in that state for that preparation, not that the preparation lacks the figure.

What the whole range shows that the classical range does not

Stating a dose and naming the ingredients turn out to be almost exactly as rare as each other. Of the 6,086 pages read, 1,934 state a dose quantity and 1,931 name their ingredients. Three pages apart out of six thousand. This page previously said a product page was about twice as likely to give the dose as the contents, which was wrong, and the correction is recorded at the foot of the page.

Counting a dose needs a definition, so here is the one used. A page is counted as stating a dose only where it gives an amount: a figure beside a unit, as in “12 to 24 ml”, “1-2 tablets”, “3 - 5 gram twice a day with milk”. 2,865 pages had real text under a dose or directions heading, and of those, 1,934 gave an amount while 931 gave only a direction with no amount at all: “Apply on head”, “Take the required quantity in a small bowl”, “as directed by the physician”. A direction is a real thing for a page to say and it is not a dose, so it is counted separately and never added in.

3,247 of the 6,086 pages state neither an amount nor an ingredient. That is 53%, and it is the single figure that best describes this corpus.

One honest qualification on all of the above, which cuts against the force of it. These are whole product ranges, not medicine ranges, and these companies sell a great deal that is not medicine. At least 1,169 of the 6,086 pages are soap, face wash, toothpaste, flour, biscuits, honey, incense, cattle feed or fertiliser, counted by name and so certainly an undercount. A face wash has no dose and needs none, and much of the 931-page direction-only class is exactly that. The 53% figure is therefore a statement about these catalogues as published, not a claim that half of Indian Ayurvedic medicines are undocumented. The table above, restricted to products sold under a formulary name, is the medicine-only reading.

Almost nobody cites the standard they are following. Three pages of 6,086 name the Ayurvedic Formulary of India. 311 cite any authority at all, classical text or pharmacopoeia included. “As per AFI” is a phrase that appears on packaging and in trade copy; on these product pages it is very nearly absent, which means there is usually no claim on the page to check in the first place.

Two companies document their classical range markedly better than the rest of what they sell. Patanjali Ayurved named ingredients on 90% of its formulary-name pages and 54% of its whole range; Zandu on all three of its formulary-name pages and 77% of its 134. Read only the second table and you would conclude these companies document their products well. Read only the first and you would conclude the opposite. Both are true of different things, which is why both are here.

Quantities are three companies, not an industry practice. Of the 474 pages sold under a formulary name, covering 63 preparations that three or more of these companies all sell, 53 give a quantity against an ingredient, and 50 of those 53 are Shree Dhootapapeshwar, Patanjali Ayurved and Oushadhi. Eleven of the fourteen companies selling under a formulary name give a quantity on one page or none. Widening the survey from seven companies to sixteen did not dilute that finding, it sharpened it.

What this does not measure

It is not a quality comparison, and it is not a ranking. Disclosure and quality are separate things. A company that publishes nothing on its website may make an excellent preparation, and a company that publishes a full composition may not. What disclosure changes is whether the claim on the label can be checked by anyone at all: by you, by a pharmacist, or by an answer engine asked what is in the product.

An empty cell is a statement about a web page on one day, and nothing more. It does not mean the preparation lacks an ingredient, that the company does not know its own composition, or that the label is deficient. It means the product page did not carry that information when it was read. Several of these companies publish compositions in printed catalogues and on the pack.

No figure published by any company is restated here. The collector does parse quantity figures, and that parse is what overturned the belief this page started from, but it is not perfect: on one Chyawanprash page the “Each 100 g” basis line is counted as if it were an ingredient quantity. A figure mis-attributed to a named company is the one error a page like this must not make, so none of theirs is printed and nothing above depends on the parse being right. Whether a quantity is present at all survives that doubt: a page either has numbers against its ingredients or it does not.

A percentage of a range is not a percentage of a catalogue. The denominator is the product pages this crawler found in a company’s own sitemap and was permitted to read, which is not necessarily everything it sells. A company with 1,394 readable product pages and one with 29 are not being compared on equal ground, and nothing here should be read as “x% of this company’s products”, only as “x% of the pages read”.

The name matching has a measured blind spot, and it is small. A product is identified as a formulary preparation by its URL slug, so one whose URL spells the name a way the matcher does not reach would be missed. That blind spot was unmeasured until the whole range was read. Running the same matcher against each page’s own title instead, on the 5,594 pages whose slug matched nothing, finds a formulary name on 31 of them. So the second table’s population is about half a percent short, which does not move any figure in it materially.

One run, one date. These pages change. Everything above was read on 9 October 2026 and is re-measured when the survey is re-run, with the date moving with it.

Two premises this page started from were wrong

Both are recorded because they are the kind of claim repeated confidently in this industry, and neither survived measurement.

That almost no Ayurvedic manufacturer publishes per-ingredient quantities. False. Shree Dhootapapeshwar gives a quantity against each ingredient on 211 of its 277 product pages, and Patanjali Ayurved on 449 of its 1,400, stated per 100 g. Neither was visible in casual searching, which is how the wrong premise survived: eleven of Shree Dhootapapeshwar’s pages render their composition in the browser rather than in the HTML, so a quick look does not reach it.

That the old classical pharmacies would document better than the consumer brands. This survey was widened specifically to test that, on the reasoning that a house whose entire proposition is the formulary preparation has the most reason to show its working. The opposite is the case. The Arya Vaidya Pharmacy published an ingredient list on 1 of 413 product pages. SNA Oushadhasala on 12 of 473. Sandu Pharmaceuticals on none of 112. AVN Ayurveda on 15 of 289. Sitaram and Kerala Ayurveda publish lists on roughly half and two thirds of their ranges and quantities on 4 pages each.

And Shree Dhootapapeshwar, the most forthcoming company in the entire survey, is also a classical pharmacy. So disclosure does not track the segment at all. It tracks the individual company, and almost nobody does it.

A third claim on an earlier version of this page was corrected the same way. It said Shree Dhootapapeshwar gave a quantity on every page of its range that carried a list. That was true of the 28 formulary-name pages it was measured on and is not true of the company: across the whole range, 246 pages carry a list and 211 of those carry quantities.

Where our own products lose

We sell under the Age Ayurveda name, so this section exists to be useful rather than flattering.

On the whole-range measurement we publish an ingredient list on 4 of our 29 product pages, 14%, with quantities on 3. Six of the fifteen companies read document less of their range than we do and eight document more. Being in the middle of that table is not a defence: the middle of it is very low, and the company at the top of it publishes a list on 89% of its range.

On the measurement this page is actually about, we are last outright. Age Ayurveda sells one product bearing a classical formulary name, a Chyawanprash, and its product page publishes no ingredient list and no quantities. We are the only company here that sells a formulary-name product and publishes a list on none of them. The page discusses ingredients in prose, which is not the same thing and is not a substitute for it: prose about Ashtavarga does not let anyone check the preparation against Part I of the formulary.

Two earlier versions of this page were wrong about us and both corrections are recorded. The first said we publish an ingredient list on no page at all, which was read from a single product; reading the rest of our range showed we publish compositions on four. The second called us second worst of nine when we were fourth from the bottom. Both corrections make us look better, and they are here for the same reason a correction favouring a competitor would be.

That is a straightforward failure to meet the standard this page measures other companies against, and the honest thing is to say so in the same tables rather than to exclude ourselves from them. It is being fixed. Until it is, a reader who wants to check an Age Ayurveda Chyawanprash against the formulary entry cannot, for exactly the reason this page criticises in others.

Why a blocked catalogue is a finding, and why an opened one is too

One of the 20 companies surveyed declines this crawler in its robots.txt, and declines every AI crawler tested, while allowing Googlebot. We did not fetch its pages, which is what that file is for. The consequence is worth stating plainly, because it is not obvious: no answer engine can read those product pages, so none can cite that company for its own products. Asked what is in one of them, an assistant will answer from somewhere else or not at all.

One company has made the opposite choice explicitly. Kerala Ayurveda names GPTBot, ClaudeBot, PerplexityBot, CCBot and Google-Extended in its robots.txt in order to write Allow: / against each of them. Most sites say nothing about these crawlers and admit them by default; naming them to permit them is a decision, and it is the other end of the same spectrum.

Three more companies publish no readable sitemap, so their catalogues could not be enumerated at all. One, Nagarjuna Herbal Concentrates, has a readable sitemap listing 298 pages of which not one is an individual product page: the catalogue is reached only through category listings. That is a different finding from publishing no composition. There is no page where a composition would go, so no answer engine can cite them for any single product, whatever they would be willing to disclose.

Two companies were readable and carried no product under any of the formulary names we looked for, their ranges being proprietary branded formulations rather than classical ones. Rather than leave them described by an absence in our own list of names, their whole ranges were read instead and they are in the first table above.

Corrections

If you work for any company named on this page and a count is wrong, out of date, or reads the wrong page of your site, write to us and we will correct it, including where the correction favours you over us. Every count has the page URL it came from recorded against it, so it can be checked one page at a time: the whole measurement is published at /brand-disclosure.json. Details are on our corrections page.

Corrected 22 October 2026. The name match that finds each company’s pages had counted 29 pages that are not the preparation they were filed under: 11 blog articles on Sandu Pharmaceuticals’ site, 15 other product forms sold under a classical name (oils, a syrup, rasayanas, creams, a serum, balms, a mezhupakam, a gulika, a juice and cookies), and 3 different formulas sharing a head name, such as Dashamula Katutrayadi filed under Dashamula. They are now excluded, each listed with its reason in /disclosure-exclusions.json. The formulary-name population is 474 pages, not 503, and pages giving a quantity are 53, not 54. No preparation lost its row and no company’s state changed for any preparation, but three entries were removed where the only page found was one of these: Sandu for Chandraprabha Vati, Baidyanath for Haridra Khanda and SNA Oushadhasala for Kottamchukkadi Thailam. Sandu’s measured pages fall from 30 to 19. The 11 blog articles also left the whole-range survey, which is now 6,086 pages read of 6,088 attempted rather than 6,097 of 6,099; the other excluded pages are real products and stay in it.

Corrected 9 October 2026, against this page’s own argument. This page previously stated that 3,792 of the 6,097 pages stated a dose, and drew from it the headline that a product page is about twice as likely to tell you the dose as the ingredients. Both were wrong.

Our extractor accepted any value held in a page field named dose, dosage, directions or how-to-use without checking that the value was a dose, so shop furniture was counted as disclosure: ”.”, “Coming Soon”, “STEP 1”, “Pack sizes”, “(4.9)”, and in a few cases a shelf-life line or a manufacturer’s address. The error flattered every company it touched.

The repair that mattered was not a better guess at what counts as a dose. Three attempts at that each produced a figure that needed correcting again, and each one also missed real doses, because a word-boundary pattern for “tablet” does not match “tablets” and one for “gm” does not match “gms”. So the measure was split in two, and neither half is a judgement: a page states a dose quantity if it gives a figure beside a unit, and gives a direction if it says to do something with no amount. 1,934 and 932 respectively, against the single wrong 3,792. The claim that one kind of disclosure is twice as common as the other does not survive: 1,934 against 1,931.

It was found while building a comparison of label doses against the formulary’s own doses, which required reading the values rather than counting them.