# Tamats — full fact sheet

Source of record: https://tamats.org
Regulatory corpus compiled: 2026-07-21
Document generated from the live corpus; counts below are computed, not typed.

## Summary

Tamats is the operating system for cosmetic product development, for brands and contract manufacturers — from first question to launch. A cosmetics-specialized AI agent searches scientific literature with cited results, analyzes ingredients, and builds the versioned formulations, compliance dossiers, stability studies and manufacturer quotes that make up the project record, checking compliance across regulatory markets and retailer clean standards along the way. Manufacturer outreach is drafted for a human to confirm before it sends.

Category: Cosmetic product development operating system.
Positioning: The operating system for cosmetic product development, where every regulatory claim carries the regulation behind it.

## What Tamats is not

- Not manufacturing execution, ERP or LIMS — no batch or lot genealogy, no production scheduling, and no supplier catalog of raw-material tradenames or certificates of analysis; raw materials are what you enter yourself.
- Not a testing laboratory — it does not run stability, challenge or clinical testing.
- Not a regulatory consultancy and not a substitute for a qualified safety assessor's report (EU CPSR, US MoCRA safety substantiation).
- Not a consumer app — it is B2B software for brand and manufacturer teams.

## Agents

All 7 agents are included on every paid plan. Tamats does
not charge per agent or per module.

- Research Agent
- R&D Agent
- Compliance Agent
- Designer Agent
- Marketing Agent
- Sourcing Agent
- Market Intel Agent

## Pricing

- Free: free, 1 seat. 25 chat messages per day, 5 images per month, 1 product and 1 formulation. No credit card, Google sign-in.
- Individual: $200 per month. Single seat, for solo founders and consultants. All seven agents, compliance dossiers, studio export. 10 dossiers and 100 images per month.
- Teams: $550 per member / month. Up to 15 members, for the whole org. Unlimited compliance dossiers, regulatory watch, and monthly limits pooled across every seat.
- Enterprise: custom pricing. Unlimited members, SSO, DPA, security review, SLA, invoicing.

Annual billing saves 17%. On Teams the
monthly limits multiply by subscribed seats and pool across the org. Sign-in is
a Google account; the free tier takes no credit card.

## Regulatory coverage

781 compiled rules total: 14 regulatory markets and
13 retailer clean standards, plus 462
ingredient pages — one for every distinct INCI term the corpus has a cited rule
about. Every entry links its primary source.

### Markets (14)

- European Union — 41 compiled entries; authority: European Commission — Regulation (EC) 1223/2009, Annexes II/III/V; page: https://tamats.org/regulations/eu
- United States — 15 compiled entries; authority: FDA — FD&C Act + MoCRA; page: https://tamats.org/regulations/us
- South Korea — 18 compiled entries; authority: MFDS — Cosmetics Act ingredient standards; page: https://tamats.org/regulations/kr
- Japan — 13 compiled entries; authority: MHLW — Standards for Cosmetics; page: https://tamats.org/regulations/jp
- China — 18 compiled entries; authority: NMPA — Safety and Technical Standards for Cosmetics; page: https://tamats.org/regulations/cn
- United Kingdom — 23 compiled entries; authority: OPSS — UK Cosmetics Regulation; page: https://tamats.org/regulations/uk
- Canada — 29 compiled entries; authority: Health Canada — Cosmetic Ingredient Hotlist; page: https://tamats.org/regulations/ca
- Australia — 5 compiled entries; authority: AICIS / TGA; page: https://tamats.org/regulations/au
- India — 24 compiled entries; authority: CDSCO / BIS — IS 4707; page: https://tamats.org/regulations/in
- Brazil — 23 compiled entries; authority: ANVISA — RDC resolutions; page: https://tamats.org/regulations/br
- Mexico — 30 compiled entries; authority: COFEPRIS; page: https://tamats.org/regulations/mx
- ASEAN — 38 compiled entries; authority: ASEAN Cosmetic Directive — Annexes; page: https://tamats.org/regulations/asean
- California (US) — 26 compiled entries; authority: OEHHA Prop 65 · AB-2771 (PFAS); page: https://tamats.org/regulations/us-ca
- Washington State (US) — 34 compiled entries; authority: Toxic-Free Cosmetics Act (HB 1047); page: https://tamats.org/regulations/us-wa

### Retailer clean standards (13)

- Clean at Sephora — 22 compiled entries; authority: Sephora clean-seal ingredient criteria; page: https://tamats.org/regulations/sephora-clean
- Conscious Beauty at Ulta — 13 compiled entries; authority: Ulta Beauty made-without list; page: https://tamats.org/regulations/ulta-beauty
- Target Clean — 13 compiled entries; authority: Target Clean icon ingredient criteria; page: https://tamats.org/regulations/target-clean
- Amazon listing compliance — 25 compiled entries; authority: Amazon restricted-products / clean criteria; page: https://tamats.org/regulations/amazon-clean
- Credo Clean Standard — 21 compiled entries; authority: The Credo Clean Standard dirty list; page: https://tamats.org/regulations/credo
- Whole Foods Body Care — 38 compiled entries; authority: Whole Foods Market body-care quality standards; page: https://tamats.org/regulations/whole-foods
- Whole Foods Premium Body Care — 31 compiled entries; authority: Whole Foods Market premium body-care standards; page: https://tamats.org/regulations/whole-foods-premium
- Walmart Sustainable Chemistry — 8 compiled entries; authority: Walmart sustainable chemistry commitment; page: https://tamats.org/regulations/walmart
- Walmart Made Without List — 176 compiled entries; authority: Walmart made-without-list substances; page: https://tamats.org/regulations/walmart-mwl
- CVS store-brand chemicals policy — 31 compiled entries; authority: CVS Health chemicals of concern policy; page: https://tamats.org/regulations/cvs
- Costco vendor compliance — 21 compiled entries; authority: Costco wholesale vendor requirements; page: https://tamats.org/regulations/costco
- Douglas clean criteria — 13 compiled entries; authority: Douglas (EU) clean beauty criteria; page: https://tamats.org/regulations/douglas
- The Detox Market banned list — 32 compiled entries; authority: The Detox Market banned-ingredient list; page: https://tamats.org/regulations/detox-market

## Free, ungated tools (no account, no API key)

- Ingredient regulatory index: https://tamats.org/ingredients
- Regulations index: https://tamats.org/regulations

## How claims are grounded

Cite-or-abstain, enforced in code rather than requested in a prompt. Factual,
safety and regulatory claims must link a real source retrieved from the corpus;
with no source retrieved the agent says so and stops. Compliance findings are
computed against the compiled rule set rather than generated as prose, and an
ingredient with no matching rule returns "review" — never a fabricated pass.
AI output is informational and is not professional, medical, legal or
regulatory advice.

## Data handling

Content submitted to the product is sent to the large-language-model provider
for inference and to any retrieval source the user invokes, under the
subprocessor terms in the DPA (https://tamats.org/legal/dpa). Private customer
content is not used to train Tamats' own models.

## Platform

Every capability, with where it stops. An ingredient list without its
boundary is a bigger claim than the product makes for itself, so the limits
are included here rather than left to the page.

### EU PIF dossier

Tamats assembles EU Product Information File Part A — composition, physicochemical and microbiological specification, stability, allergen and undesirable-effects sections — from the formulation, stability studies and cited literature already held in the workspace, and tracks each section through empty, agent-drafted and human-confirmed states. It does not write Part B: the safety assessment and the Cosmetic Product Safety Report require a qualified safety assessor, and the tool refuses to draft one under any phrasing, marking those sections assessor-only instead. The output is a dossier a real assessor can pick up with the ingredient-level regulatory work already done and cited, not a filing you can submit as-is.

Limits:
- No Part B. The safety assessment and the CPSR are not drafted, summarised or approximated — the tool refuses the request rather than producing something that reads like one.
- No toxicological calculation. There is no margin-of-safety, NOAEL or SED computation anywhere in the product.
- Part A only, and only for the EU. There is no equivalent dossier engine for Korean, Japanese or Chinese registration.
- No electronic submission. Nothing files to CPNP; the output is a dossier, not a transmission.

Page: https://tamats.org/platform/eu-pif-dossier

### Compliance screening

Tamats screens ingredients and claims against a compiled corpus of government regulations and retailer clean-beauty standards, and every finding links the primary source it was compiled from. Findings are computed against the rule set rather than generated as prose, so an ingredient with no matching rule returns 'review' — an honest gap — instead of a fabricated pass. The corpus is a screen, not a legal opinion: it tells you what to look at and shows you the text, and a regulatory professional still makes the call on a borderline case.

Limits:
- A screen is not a regulatory opinion. It narrows what needs a professional's attention; it does not replace the professional.
- An ingredient with no matching rule returns 'review'. That is a gap in the corpus being reported honestly, not a clearance.
- Coverage is what has been compiled from primary sources. Markets outside the compiled set are not silently approximated from adjacent ones.
- No safety assessment. Screening an ingredient against a limit is a different thing from assessing a product's safety, and only the first happens here.

Page: https://tamats.org/platform/compliance-screening

### Cited literature search

Tamats searches PubMed for peer-reviewed studies and the open web — general, scholar and news — and returns results with the title, source and link attached, never a claim with nothing behind it. A search that finds nothing says so and stops rather than filling the gap with a recalled fact, and any paper worth keeping saves to the organisation's shared library with its citation intact. PubMed is the tool for evidence; general and news web search cover trend and regulatory-change questions PubMed was never built to answer, and the agent is instructed to reach for the right one.

Limits:
- PubMed and web search only — there is no full-text access to paywalled journals; a search returns the abstract and a link, not the paper itself.
- A search with no PubMed hits abstains rather than answering from the model's own memory of the literature — that abstention is a finding, not a failure to fix.
- Saving a paper stores the citation and metadata your team chose to keep, not the corpus itself — Tamats does not host or redistribute journal content.

Page: https://tamats.org/platform/cited-literature-search

### Market & social intelligence

Tamats searches public posts and videos on TikTok, Instagram, Xiaohongshu, YouTube and Reddit for a topic and returns ranked results with engagement metrics and source links, alongside general, news and shopping web search for market and competitor signal. It ships one standing monitor kind end to end — a competitor-launch keyword watch that a daily cron runs and reports into the workspace — not a general monitor builder for arbitrary conditions. Social and market signal is explicitly lower citation authority than PubMed literature or the compiled compliance corpus: it is directional evidence for a brief, never the source behind a safety or regulatory claim.

Limits:
- Exactly one monitor kind ships end to end — competitor-launch keyword watches. There is no monitor builder for price changes, stock levels or arbitrary conditions yet.
- Social search carries no citation authority for a factual, safety or regulatory claim — it is explicitly excluded from that path and is directional only.
- The social search platforms cannot all be searched the same way — a platform with no reliable keyword-search endpoint is left out of discovery entirely rather than offered and silently returning nothing.
- Social search needs a configured key and a funded account; an unconfigured or depleted account returns an honest error, not an empty 'no results'.

Page: https://tamats.org/platform/market-and-social-intelligence

### Formulation workspace

Tamats holds a cosmetic formulation as phases of ingredients with percentages, checks the running total against 100% as you build, and versions every save as a semantic-version revision so a reformulation is a new number, not an overwritten row. Measured bench results — pH, viscosity, SPF, a pass/fail/pending stability call — log against the formula directly, and comments thread on the record itself rather than living in a separate chat. The INCI label generates from the same phases, sorted by descending concentration with the sub-1% ordering rule applied, matched to the target region's labelling conventions but never a substitute for a human proofing the final artwork.

Limits:
- Balance checking is arithmetic, not chemistry — a formula that sums to 100% is not thereby verified as stable, compatible or safe.
- The INCI label generator applies the descending-order and 1% line rules; it does not check whether the listed ingredients are legal in any market — that is a separate compliance screen.
- A test log is a place to record a measurement your own bench or a lab already took. Tamats does not run any test itself.

Page: https://tamats.org/platform/formulation-workspace

### Ingredient intelligence

Tamats resolves a single ingredient — by common or INCI name — against a CosIng-derived dictionary for its function and CAS/CosIng identity, a live CIR safety search for its cited safety record, a curated set of pairwise incompatibility rules, a check for whether it sits at or below the INCI 1% ordering line as a marketed active (underdose, or 'fairy-dusting'), and the EU Annex III fragrance-allergen declaration list. Every field says where it came from, and an ingredient with no CIR record or no matched incompatibility rule reports that absence honestly rather than implying safety. It has no molecular structure data behind it, so there is no structural-similarity or QSAR prediction, and no comedogenicity rating anywhere in the product.

Limits:
- No molecular structure data — there is no SMILES string, structural-similarity search or QSAR prediction anywhere in the product.
- No comedogenicity rating. That dataset does not exist here and nothing approximates it.
- Incompatibility coverage is a small, curated set of known pairwise conflicts, not a general cosmetic-chemistry safety model — a clean result means nothing in that set matched, never that the formula is compatible.
- Underdose detection only fires when a concentration or a list position can place the active at or below the 1% line; with no signal at all it abstains rather than guessing.

Page: https://tamats.org/platform/ingredient-intelligence

### Formula reverse engineering

Tamats takes a product name, a raw INCI list, a product page URL, or a photo of a back label and turns it into a per-ingredient percentage range, a confidence tier, and a deterministic phase grouping — a bench starting point, never a finished formula or a point-estimate composition. The ranges come from descending-order labelling rules and compiled legal maxima, not a lab assay: no product was tested, so the true composition can differ from the estimate at any point the underlying INCI order doesn't legally constrain. A bare product name with no list, URL or photo is recalled by the language model and is explicitly marked unverified, with no citations attached.

Limits:
- This is inferred from label order and legal ceilings, not an assay — no chromatography, no lab test, nothing measured on the actual product. Treat every range as a hypothesis to confirm on the bench, not a verified composition.
- A bare product name with no list, URL or photo produces an unverified result recalled by the language model, with no citations — always prefer the real INCI list, the product page, or a label photo.
- Per-ingredient bands are never rescaled to force their midpoints to sum to 100% — doing so would manufacture precision the method doesn't have, so the sum is reported as an honest disclosure signal instead.
- Below the 1% ordering line, position stops being legal evidence — confidence drops to low there by design, and the lower bound widens rather than staying tightly guessed.

Page: https://tamats.org/platform/formula-reverse-engineering

### Stability studies

Tamats schedules a cosmetic stability study — conditions like accelerated, ambient, cold, freeze/thaw cycling and light exposure, each with its own timepoint checkpoints — and generates the concrete due dates for every reading up front. Each observation logs pH, viscosity and appearance against that timepoint, and is judged pass, fail or drift against the spec bounds you set, with drift computed from the trend across observed points rather than a single reading. Tamats does not run the testing itself: a lab or your own bench places the samples and takes the measurements, and there is no shelf-life prediction model behind any of it — the tool is the schedule and the record, not the chamber.

Limits:
- Tamats does not run the study. A lab or your own bench places the samples in the actual conditions and takes the actual measurements — this is the schedule and the record over that work, not a replacement for it.
- There is no shelf-life or expiry prediction model. A pass/fail/drift verdict describes one observed reading against your spec, not a forecast of when the product will fail.
- Drift detection needs several observed points in a series to call a trend honestly — with too few, it reports no drift rather than guessing from too little data.

Page: https://tamats.org/platform/stability-studies

### US MoCRA status

Tamats computes a US MoCRA small-business exemption verdict from the organisation's own sales band and any statutory exclusion category, tracks each product listing's status against a full safety-substantiation checklist, and logs adverse events with the statutory reporting due date computed automatically from the date received. Every field is read from the organisation's own saved MoCRA record — an unset sales band returns an honest 'cannot be determined' rather than a guessed exemption. Nothing here files with the FDA: facility registration, product listing submission and adverse-event reporting all stay outside the product, and every rule it applies links out to the primary FDA source behind it.

Limits:
- No electronic submission to the FDA — nothing here files a facility registration, a product listing, or an adverse-event report. It computes what those filings would need and links to where you actually file them.
- The exemption verdict comes back as 'cannot be determined', not a guess, when the sales band is unknown — this tool will not assume exempt or not-exempt from missing data.
- Safety substantiation is a checklist of what evidence exists, not the evidence itself or a judgement that it's adequate — that judgement stays with whoever files the listing.

Page: https://tamats.org/platform/us-mocra

### Regulatory watch

Tamats raises a regulatory-watch alert only when a revision to the compiled compliance corpus changes an ingredient's status or limit in a way that touches an ingredient already in one of the organisation's own formulations, or a market or retailer standard the organisation watches. Each alert names the ingredient, exactly what moved — a status change, a new or tightened concentration limit, a rule added or removed — the region or retailer scope, the affected formula, and the primary source behind the change. An empty result is reported honestly and distinguishes no open alerts from the changelog holding only a baseline snapshot; it is never presented as a clean bill of health.

Limits:
- Alerts only fire on rules already compiled into the corpus — a market or retailer standard outside that compiled set produces no alert, silently or otherwise, because there is nothing to diff.
- This watches the compiled ruleset, not primary regulator publications directly — a change has to reach the corpus before it can raise an alert here.
- An alert is a notice to review, not a verdict — it does not itself reformulate or resubmit anything; a human still decides what the change means for the affected product.

Page: https://tamats.org/platform/regulatory-watch

### Manufacturer sourcing

Tamats keeps a record of manufacturer contacts, generates a structured RFQ email listing every formulation ingredient with its INCI, target percentage, quality grade and MOQ ask, and drafts sample-request or general outreach the same human-in-the-loop way — every draft is exactly that, a draft, and nothing sends until a person clicks send. Sending itself needs a mailbox the organisation has connected (Google or Microsoft); without one, a draft still gets written and simply waits, and it never falls back to a shared platform sender that would put a reply in the wrong inbox. Inbox search reaches threads already linked to a manufacturer or influencer contact — mail outside that scope is invisible to it, not absent.

Limits:
- Nothing sends without a human clicking send — every RFQ, sample request or outreach draft is confirm-before-send, with no exception.
- Sending needs a connected mailbox (Google or Microsoft). Without one, a draft is still written and waits — it never sends through a shared platform address instead.
- Inbox search reaches only threads already linked to a manufacturer or influencer contact; mail outside that scope is not searched, and a miss there is not evidence the mail doesn't exist.

Page: https://tamats.org/platform/manufacturer-sourcing

### Quote intake

Tamats parses a manufacturer's quote — pasted text or an existing inbox message — into structured commercial terms: unit price and currency, MOQ, lead time, incoterm and a validity date, using plain pattern matching rather than a model's guess. A term the parser can't find stays absent and is listed as unparsed rather than defaulted to a plausible-looking number, and every parsed quote sits in a parsed state until a human confirms or rejects it — nothing here treats a parse as a decision. Currency is never converted; a comparison across quotes in different currencies is flagged as such rather than silently normalised.

Limits:
- A field the parser cannot resolve is left absent and listed as unparsed — it is never defaulted, guessed, or filled from a typical value for the category.
- Nothing here persists automatically. A human confirms the parsed fields in the quote UI before they count as the organisation's record of the quote.
- Currency is never converted — a comparison across quotes in different currencies is flagged as mixed and incomparable rather than normalised to one currency for you.

Page: https://tamats.org/platform/quote-intake

### Packaging & label studio

Tamats builds a three-dimensional vessel from a small set of parametric shape archetypes — cylinder, tapered, shouldered, jar and others — with a print-area label zone, computes the flat dieline that wraps it from the vessel's own dimensions, applies your brand kit's palette, and assembles the whole thing into a tech pack ready for print export. There is no supplier packaging catalogue behind any of it: every vessel and stock component in the studio is entered by your own organisation, because an earlier fabricated seed catalogue was deliberately removed rather than left in place implying suppliers that don't exist. The dieline and tech pack are computed geometry and assembled data, not a rendering service or a supplier quote.

Limits:
- There is no supplier packaging catalogue. Every vessel and stock component in the studio is entered by the organisation itself — an earlier seed catalogue of fabricated supplier components was deliberately deleted rather than left in place.
- The dieline is computed geometry over the dimensions entered, not a validated die-cut file — a converter should confirm it against their own tooling before cutting.
- There is no rendering or photography service behind the vessel view — it is a parametric geometry preview, not a photorealistic product shot.

Page: https://tamats.org/platform/packaging-and-label-studio

### Connected sources

Tamats searches an organisation's connected S3 documents live and keyword-matches against them with nothing copied into Tamats' own storage, answers questions against a connected research database through a guarded text-to-SQL layer that only ever generates a single read-only SELECT statement, and connects to a Shopify or Cafe24 storefront to read the merchant's own live catalogue and orders. Every one of these is read-only and zero-store: disconnect the source and Tamats has nothing left of it, because nothing was ever kept. An organisation with nothing connected gets an honest 'no source connected' rather than a demo dataset standing in for one.

Limits:
- Zero-store means exactly that — Tamats does not index, cache or persist a copy of any connected S3 document, database row, or storefront record between queries.
- The text-to-SQL layer only ever generates a SELECT statement against a read-only role — it cannot write, alter or delete anything in the connected database, by design, not just by convention.
- Shopify's admin tier and Cafe24's admin tier both need the organisation to complete an OAuth connection before write-adjacent data (orders, inventory) is reachable — the storefront tiers alone only see what's already public.

Page: https://tamats.org/platform/connected-sources

### Distribution & creators

Tamats tracks the organisation's distribution channels through a staged pipeline — from first conversation to live shelf — with buyer contacts and notes on each channel, and separately builds a creator roster from a plain-language brief, enriching gaps like gender (a labelled heuristic, never asserted as fact), contact and engagement metrics. A monitored campaign post's view, like and comment counts are re-checked over time from the platform itself, and any outreach draft — to a channel contact or a creator — is a human-confirmed send, never an automatic one. Comment counts are public and trackable; comment bodies are not read or stored.

Limits:
- Gender and aesthetic scoring on a creator profile are labelled heuristics, not verified facts — they are shown with that label, never asserted as certain.
- Comment bodies on a monitored post are never read or stored, only the public comment count — anything said in a comment stays on the platform.
- Creator discovery only searches platforms where the underlying search actually works by keyword; a platform with no reliable search endpoint for this is left out of the picker rather than offered and silently returning nothing.
- Nothing sends without a human clicking send from a connected mailbox — an outreach draft with no mailbox connected simply waits, unsent.

Page: https://tamats.org/platform/distribution-and-creators

### Post-launch feedback

Tamats classifies imported customer feedback — adverse reaction, efficacy, texture, scent, packaging, shipping, price, or unknown when nothing in the lexicon matches — using a deterministic, bilingual keyword lexicon rather than a model's judgement call, and adverse-reaction terms always win the category and are always flagged high severity, on the principle that under-calling a reaction is the expensive mistake. A per-product, per-category rollup flags a spike when recent volume clears a floor and genuinely rises against the equivalent prior window, and a feedback item flagged as an adverse reaction can be routed straight into the same dossier section that assembles undesirable-effects data for an EU PIF. Ingestion is CSV import only today — the feedback schema already has room for a direct Shopify or Cafe24 pull, but nothing wires it up yet, so a store's own reviews still have to be exported and imported by hand.

Limits:
- Ingestion is CSV import only. The schema already has a place for a direct Shopify or Cafe24 pull, but nothing automatically ingests from those connectors yet, even though both exist as connected sources elsewhere in the product.
- Classification is a keyword lexicon, not sentiment analysis or a model judgement call — an item with no matching term is marked unknown rather than guessed at, and a sarcastic or ambiguous review can land in the wrong bucket the same way any keyword match can.
- Spike detection needs a minimum recent count before it reports anything at all — a small absolute jump in a low-volume category is deliberately not flagged as a spike.
- Routing a reaction into the dossier drafts a section for a human to review; it does not itself notify a safety assessor or trigger any regulatory filing.

Page: https://tamats.org/platform/post-launch-feedback

Hub: https://tamats.org/platform

## Comparisons

- vs Coptis: A long-established cosmetics PLM suite, sold to R&D laboratories and manufacturers running formal product-development cycles. https://tamats.org/vs/coptis
- vs Trace One Devex PLM: Enterprise PLM for consumer packaged goods, including cosmetics and personal care, sold into large brand and retailer organisations. https://tamats.org/vs/trace-one-devex
- vs Good Face Project: The nearest neighbour in category: cloud formulation software built around a proprietary ingredient ontology, sold across the whole value chain. https://tamats.org/vs/good-face-project
- vs CM Studio+: A free-tier formulation platform bundled with an ingredient marketplace — the closest thing in the category to a zero-cost starting point. https://tamats.org/vs/cm-studio-plus
- vs Formulator Online: Affordable, published-price formulation software for indie makers and small brands — the practical alternative for a one-market launch. https://tamats.org/vs/formulator-online
- vs Cosmetri (Registrar Corp): Cosmetics PLM plus GMP software, sold alongside a regulatory services business — software and the humans who file, in one contract. https://tamats.org/vs/cosmetri
- vs Worldover: An enterprise consolidation play: one AI-orchestrated system intended to replace the whole operational stack for substance-based businesses. https://tamats.org/vs/worldover
- vs ChatGPT and other general AI assistants: The tool most cosmetic teams actually use for regulatory questions today — and the one that quietly produces the most expensive mistakes. https://tamats.org/vs/chatgpt
- vs spreadsheets and a regulatory consultant: The default stack for most indie and scaling brands: a formula workbook, a folder of PDFs, and a consultant on retainer or per project. https://tamats.org/vs/spreadsheets-and-consultants

Hub: https://tamats.org/alternatives

## Questions and answers

Q: What is Tamats?
A: Tamats is the operating system for cosmetic product development — for cosmetic brands and contract manufacturers, from first question to launch. You chat with a cosmetics-specialized AI agent that searches scientific literature with cited results, analyzes ingredients against a compiled regulatory corpus, and builds the versioned formulations, compliance dossiers, stability studies and manufacturer quotes that stay in the workspace as the project's record. Every regulatory claim carries the regulation behind it, and manufacturer outreach is drafted for a human to confirm before it sends.

Q: Who is Tamats for?
A: Two sides of the same deal: cosmetic brands (indie through scaling) who sell and outsource manufacturing, and contract manufacturers and ODMs who formulate and produce. Within those teams the daily users are product development, regulatory affairs, and founders who wear both hats.

Q: How much does Tamats cost?
A: There is a free tier with one seat and no credit card. Paid plans are Individual at $200 per month for a single seat, and Teams at $550 per member per month for up to 15 members, with annual billing saving 17%. Enterprise is custom-priced. Every paid plan includes all seven agents — Tamats does not charge per agent or per module — and on Teams the monthly limits pool across every seat.

Q: Is there a free version of Tamats?
A: Yes. The free tier gives one seat with a Google sign-in and no credit card: 25 chat messages per day, 5 images per month, and one product and one formulation. Compliance dossiers are a paid feature. Separately, two tools are ungated and need no account at all: the ingredient regulatory index at tamats.org/ingredients, and the regulations index at tamats.org/regulations.

Q: Which regulatory markets does Tamats cover?
A: Fourteen markets: the European Union, United States (FD&C Act and MoCRA), South Korea, Japan, China, United Kingdom, Canada, Australia, India, Brazil, Mexico, ASEAN, plus California (Prop 65 and AB-2771) and Washington State (Toxic-Free Cosmetics Act) as separate US sub-markets. Each rule links the primary source it came from.

Q: Does Tamats check retailer clean standards like Clean at Sephora?
A: Yes. Thirteen retailer programmes are compiled alongside the government regulations: Clean at Sephora, Conscious Beauty at Ulta, Target Clean, Amazon listing compliance, Credo Clean Standard, Whole Foods Body Care and Premium Body Care, Walmart Sustainable Chemistry and the Walmart Made Without List, CVS store-brand chemicals policy, Costco vendor compliance, Douglas clean criteria, and The Detox Market banned list. Retailer exclusion lists are a common reason a compliant formula still fails to get shelf space, so they are checked in the same pass as the legal ones.

Q: How does Tamats stop the AI from inventing regulatory answers?
A: Cite-or-abstain, enforced in code rather than asked for in a prompt. Factual, safety and regulatory claims must link a real source retrieved from the corpus; when no source is retrieved the agent says so and stops instead of guessing. Compliance findings are computed against the compiled rule set, not generated as prose, and an ingredient with no matching rule returns 'review' — never a fabricated pass.

Q: How is Tamats different from asking ChatGPT or Claude directly?
A: A general assistant answers cosmetic regulatory questions from memory, which is where hallucinated concentration limits and retired annex entries come from. Tamats computes findings against a compiled corpus of regulations and retailer standards with a citation on every entry, keeps formulations as versioned objects instead of chat text, and abstains when it has no source. The tradeoff is scope: it only does cosmetics.

Q: Is Tamats a PLM system?
A: No. Formulations, compliance dossiers and stability records accumulate in Tamats as your project's own record, but it does not replace manufacturing execution, ERP, LIMS or QMS — there is no batch or lot genealogy, no production scheduling, and no supplier raw-material catalog. And it is not a testing laboratory: stability, challenge and clinical testing still happen outside Tamats.

Q: Can Tamats replace a safety assessor or a CPSR?
A: No. AI output in Tamats is informational and is not professional, medical, legal or regulatory advice. An EU Cosmetic Product Safety Report and US MoCRA safety substantiation still require a qualified assessor. What Tamats does is get a formula to that assessor with the ingredient-level regulatory work already done and cited.

Q: Can I check a formula for compliance without creating an account?
A: Checking a full formula against a market or retailer standard requires the free tier — one seat, Google sign-in, no credit card. Without an account you can still look up any single ingredient at tamats.org/ingredients or browse the rules themselves at tamats.org/regulations, both against the same compiled corpus the authenticated compliance surfaces use.

Q: What agents does Tamats include?
A: Seven, all included on every paid plan: Research (cited literature search), R&D (formulation building and versioning), Compliance (multi-market and retailer checks), Designer (concept imagery, label editing, brand kit), Marketing (positioning and claims support), Sourcing (manufacturer discovery and outreach drafts), and Market Intel (category and competitor research).

Q: Does Tamats help find manufacturers?
A: Yes. The Sourcing agent discovers candidate manufacturers and drafts outreach, including RFQ-shaped briefs built from the formulation you have in the workspace. Sending is human-confirmed — the agent never emails a manufacturer on its own.

Q: Is my formulation data used to train AI models?
A: No. Content you submit is sent to the large-language-model provider for inference and to any retrieval source you invoke, processed under the subprocessor terms in the DPA. Your private content is not used to train Tamats' own models.

Q: Does Tamats support Korean, Japanese and Chinese market entry?
A: Yes — South Korea (MFDS Cosmetics Act ingredient standards), Japan (MHLW Standards for Cosmetics) and China (NMPA Safety and Technical Standards for Cosmetics) are three of the fourteen compiled markets, each with cited ingredient rules. The marketing site is published in English, Korean and Japanese.

Q: How are Tamats usage limits metered?
A: Four things carry a monthly limit — compliance dossiers, generated images, outreach sends and social searches — and everything else is unmetered. Individual gets 10 dossiers, 100 images, 20 outreach sends and 50 social searches per month. Teams gets unlimited dossiers, and its image, outreach and social limits multiply by the number of subscribed seats and pool across the whole org, so a heavy user and a light user draw from the same balance. Tamats no longer prices in credits.

## Contact

Email: team@tamats.org
Product app: https://app.tamats.org