Generate bench-ready cosmetic formulations with built-in regulatory intelligence across 21 global markets.

theformulator.ai generates a bench-ready cosmetic formulation from a written product brief. Every ingredient concentration is checked against the regulatory limits of the markets named in your brief before the formula is returned, and each ingredient's inclusion is justified against usage data from 1M+ launched products. Output is INCI-only and supplier-agnostic: a Quick Formula plus up to three exploration variants, typically in under five minutes.

theformulator.ai helps cosmetic formulators and innovation teams turn product briefs into structured prototype formulations, with ingredient guidance, regulatory screening across 21 markets, and formulation troubleshooting built in.

Designed for professional formulators and R&D teams. Not a consumer recipe tool.

Start free, no credit card required.

Early access is a curated cohort of professional formulators, NPD and regulatory teams.

See sample output →
1M+ commercially launched products34,000+ peer-reviewed research papers429,000+ research findings extracted44,000+ ingredients with technical profilesSupplier-Agnostic
theformulator.ai
27 / 40 creditsJR

Good morning, Formulator.

3 active formulations · 2 regulatory alerts

12
FORMULATIONS
38
SAVED
2
REG. ALERTS
5
MARKETS
Brightening Serum
EUUSIN
SPF 50 Sunscreen
EUCNJP

REGULATORY ALERTS

NMPA limit update

Titanium Dioxide max. revised

EU Annex II review

3 ingredients under assessment

TODAY'S INTELLIGENCE

REGSCIENCEMARKETINGREDIENT

Built for the teams that turn concepts into products.

R&D Formulators

Generate structured starting formulations and reduce dead-end iterations with market-validated ingredient systems.

Innovation Teams

Translate product briefs and claims into formulation pathways faster, from concept to bench-ready prototype.

Regulatory Affairs

Catch ingredient and market-fit issues earlier in development, across 21 regulatory markets including Brazil, Thailand, and Malaysia.

Technical & Sales Teams

Respond to customer briefs with higher-quality prototype concepts backed by real formulation intelligence.

THE PROBLEM

Formulation development hasn't changed in decades.

Cosmetic formulators still work from disconnected PDFs, spreadsheets, and supplier portals. We've unified the intelligence.

Professional formulators still navigate regulatory requirements across PDFs and spreadsheets. Supplier data lives in disconnected portals. When something goes wrong in stability testing, you start over.

  • Regulatory limits change across 21 markets, no single source of truth
  • 1M+ commercially launched products contain co-occurrence intelligence that nobody has organised
  • CIR safety data, supplier TDS, research literature, all disconnected
  • Every failed stability batch is a reformulation from scratch
THE KNOWLEDGE BASE

One of the most comprehensive dedicated knowledge systems for cosmetic formulation and compliance.

1M+marketed products

ingredient co-occurrence intelligence across skin care, hair care, sun care, and personal care

44,000+cosmetic ingredients profiled

ingredient-level data supporting selection, screening, and compatibility checking

25,000+reference formulations

historical prototype patterns to support generation and benchmarking

98,000+cosmetic patents trained on

formulation and efficacy signal learned from the A61K8 cosmetic patent corpus, not a compliance or safety claim

scored ingredient safety profiles

peer-reviewed safety data integrated at ingredient level

34,000+peer-reviewed research papers

full-text extraction from PMC, Europe PMC, CORE, not just abstracts

429,000+peer-reviewed findings

safety, efficacy, and formulation data extracted from primary literature

21regulatory markets

EU, UK, USA, China, Japan, Korea, India, Australia, Canada, Brazil, Turkey, Taiwan, Thailand, Malaysia, and more

🇪🇺EU
🇺🇸US
🇬🇧UK
🇨🇳China
🇯🇵Japan
🇮🇳India
🇰🇷Korea
🇨🇦Canada
🇦🇺Australia
🇧🇷Brazil
🇹🇷Turkey
🇹🇼Taiwan
🇹🇭Thailand
🇲🇾Malaysia
🇸🇬Singapore
🇮🇩Indonesia
🇻🇳Vietnam
🇵🇭Philippines

Green — full regulatory coverage · Amber — coverage expanding in 2026

Sources include: 34,000+ peer-reviewed papers (PMC, Europe PMC, CORE, PubMed, OpenAlex) · CIR safety assessments · ECHA CLP/SVHC · SCCS opinions · IARC/NTP carcinogen classifications · EUR-Lex Annexes · MHLW Japan · NMPA China · MFDS Korea · MOCRA · CDSCO India

This corpus took years to assemble. It is the foundation that makes every formulation, market intelligence suggestion, and regulatory check grounded in real corpus data.

HOW IT WORKS

Brief. Generate. Test. Ship.

01

Define the brief

Enter product format, claims, target markets, sensory goals, and technical constraints across 5 structured layers.

sulphate free · no silicones · EU + India · brightening

02

Generate formulation routes

Receive structured prototypes where every concentration is pre-validated against your target markets. Regulatory limits, ingredient compatibility, and HLB requirements are checked during generation, not after.

Aqua · Glycerin · Niacinamide · Centella Asiatica Extract · Bakuchiol

03

Screen for fit and risk

Evaluate ingredient suitability, market-specific regulatory constraints, and formulation confidence indicators across 21 markets.

Stability 82% · Performance 91% · Regulatory 76%

04

Refine and troubleshoot

Compare variants, diagnose instability concerns, and iterate toward a stronger starting formula with the Formulation Partner.

emulsion break · pH drift · preservative efficacy

SEE WHAT YOU GET

Three levels of intelligence.

Choose the depth your project needs.

Quick Formula

1 credit

A bench-ready formula with regulatory pass/fail per market and preservation system. Built for speed, ideal for early-stage screening.

INCI formula table · regulatory status summary · preservation system

Download sample →
Most popular

Intelligence Brief

3 credits

Full formulation intelligence with safety scores, per-ingredient regulatory detail, market co-occurrence data, and preservation rationale. The workhorse for professional R&D.

Everything in Quick plus safety overview · market intelligence · formulation notes · ingredient-level regulatory table

Download sample →

Dossier

5 credits

Complete technical package with stability risk assessment, manufacturing brief, technical ingredient data, and PubMed citations. Structured for your regulatory file and CMO handoff.

Everything in Brief plus stability risk assessment (ICH zones) · manufacturing brief · technical ingredient data · 3–5 PubMed citations per active

Download sample →

Explore up to 3 variants of your formula, then lock the one you want: credits buy the depth of that report, not the number of formulas.

INTELLIGENCE ARCHITECTURE

How Our Formulation Engine Works

Every formulation is built from 17 specialised data layers — and the engine reasons twice before a single ingredient is committed.

01

Brief Analysis

Your formulation brief is parsed to identify product type, target markets, ingredient constraints, claims, and technical requirements. This determines which data sources activate and how deeply the engine retrieves.

02
FIRST INTELLIGENCE PASS

Intelligent Ingredient Selection

The engine doesn't pick ingredients from a dropdown. It reasons about your brief against real data, such as surfactant mildness scores, preservative pH compatibility, market co-occurrence patterns, and safety profiles, to assemble a candidate pool. Every ingredient earns its place through data, not defaults.

Market Intelligence

Co-occurrence patterns from 1M+ commercially launched products

Surfactant Properties

Mildness-ranked systems with Zein Number and clinical irritation data

Preservative Systems

Pre-validated systems ranked by spectrum, pH range, and natural origin percentage

Ingredient Hazard Data

Raw hazard classifications across ingredients from ECHA CLP, IARC Monographs, NTP, and REACH SVHC candidate lists

Ingredient Technical Data

Use levels, phase of addition, processing notes, solubility, HLB from 44,000+ profiles

COSING Reference Data

Canonical INCI names, CAS numbers, EC numbers, and regulatory functions for 30,000+ ingredients

Regulatory Intelligence

Per-ingredient status across 21 markets — prohibited, restricted, permitted, with concentration ceilings

Concentration Limits

Maximum use levels from EU Annexes, Japan Standards, Korea MFDS, India IS 4707, and more

Peer-Reviewed Literature

429,000+ extracted findings on safety, efficacy, and formulation science from 34,000+ papers

CIR Safety Assessments

Cosmetic-specific safety conclusions from 2,600+ peer-reviewed final reports

Proprietary Safety Engine

6-axis hazard scoring (carcinogenicity, reproductive toxicity, sensitization, systemic toxicity, irritation, and endocrine disruption) computed from ECHA, SCCS, SVHC, and IARC primary data with exposure-adjusted weighting

Retailer Compliance

Ingredient restriction lists for Sephora Clean, Credo, Ulta Conscious Beauty, and Whole Foods

Reference Formulations

25,000+ professional benchmark formulations with phase structure and concentration patterns

Representative layers — 17 specialised data layers feed the engine in total.

03
SECOND INTELLIGENCE PASS

Formulation Generation

The selected ingredients are assembled into per-ingredient dossiers — each containing technical data, regulatory status, safety scores, literature findings, and a pre-resolved effective concentration range. A second, independent intelligence pass receives these dossiers and designs the complete formulation: phase assignments, concentrations, processing order, and preservation strategy. It formulates with full regulatory awareness — it doesn't guess concentrations, it checks limits.

04

4-Sweep Validation

Every output passes four independent validation sweeps: prohibited substance screening, concentration compliance against regulatory ceilings, restricted substance flagging, and data-driven phase placement verification. Trade names and supplier references are stripped at three independent layers — the platform is supplier-agnostic by architecture, not by policy.

17 specialised data layers. Two independent intelligence passes. Four validation sweeps. The complete formulation — including regulatory screening across all target markets — generates in minutes.

Powered by regulatory intelligence across 21 markets, ingredient safety methodology, market intelligence from real products, and the chemistry knowledge base.

Built for how formulators actually work

Every feature is grounded in real regulatory data, real market intelligence, and real formulation science.

Compliance

The platform flags regulatory conflicts before you generate.

As you build your brief, real-time intelligence surfaces market-specific tensions. EU/China conflicts, Ecocert constraints, prohibited ingredient flags, all resolved before a single INCI is committed.

REGULATORY STATUS — 21 MARKETS
EU
US
UK
CN
JP
IN
KR
CA
AU
BR
TR
TW
TH
MY
SG
ID
VN
PH
CN — NMPA 2023 CONFLICT

Oxybenzone prohibited under CN revision. 3 compliant alternatives ranked.

Market Intelligence

Know what is already working in 1M+ products.

Co-occurrence analysis across 1M+ marketed formulations surfaces ingredient combinations that correlate with consumer acceptance in your target markets, giving every output a market-intelligence foundation.

CO-OCCURRENCE — APAC BRIGHTENING (1M+ products)
Niacinamide + Centella4,218
Niacinamide + Bakuchiol3,041
Centella + Ceramide NP2,187
Bakuchiol + Retinol891
Formulation

Phase-structured. Regulatory-validated. Ready for the bench.

INCI-only output: no brand names, no supplier bias. Phase A through C clearly structured with critical control points flagged and 3-dimensional confidence scores across stability, performance, and regulatory safety.

INCI OUTPUT — VARIANT 1 · INCI-ONLY
INCI NamePh.Conc.Conf.
AquaAq.s.
98%
GlycerinA5.00%
97%
NiacinamideA4.00%
94%
Centella Asiatica ExtractB2.00%
88%
BakuchiolB0.50%
85%
Phenoxyethanol, EthylhexylglycerinC0.80%
91%

Confidence indicators are generated using internal evaluation logic across ingredient compatibility, known formulation patterns, regulatory constraints, and category fit. They support expert review; they do not replace laboratory validation or formal safety assessment.

Platform Integrity

No supplier relationships. No commercial influence.

Every ingredient recommendation is ranked by performance data, market frequency, and regulatory status. No supplier pays to influence your formulation output. Platform neutrality is architectural.

INGREDIENT RANKING METHODOLOGY
Market frequencyStability dataRegulatory statusCIR safety
theformulator.ai
Ranked formulation output
No supplier pays to influence output
Safety

ingredient safety profiles integrated into every output.

Every ingredient is cross-referenced against safety profiles built from CIR assessments, ECHA dossiers, SCCS opinions, IARC/NTP classifications, and 34,000+ peer-reviewed research papers. Safety tiers are computed per-ingredient and flagged in the output before you reach the lab.

Read our safety scoring methodology →
SAFETY INTELLIGENCE — CIR CROSS-REFERENCE
NiacinamideSAFE
BakuchiolSAFE
OxybenzoneRESTRICTED
PhenoxyethanolCONDITIONALLY SAFE
FORMULATION PARTNER

Your AI co-pilot for formulation troubleshooting.

Paste a failing formula. Get a ranked diagnosis.

Describe an emulsion breaking at 40°C. Get three ranked root causes with probability scores, cross-referenced against thousands of similar systems in the knowledge base. Fix recommendations included.

FAILURE DIAGNOSIS — RANKED BY PROBABILITY
ROOT CAUSE #1HIGH — 78%

Insufficient HLB balance for O/W emulsion at 40°C.

FIX

Increase Cetearyl Alcohol to 2.5%. Cross-ref: 412 stable O/W systems.

ROOT CAUSE #2MED — 54%

pH drift above 6.2 destabilising carbomer network.

FORMULATION VAULT

Your Formulations Stay Private

Every formulation you generate is stored in your private workspace. No one at theformulator.ai — including our team — can access your generated formulations. Export or delete your data anytime.

Private Workspace

Your generated formulations are stored in a workspace tied to your account. No theformulator.ai team member can read your formulation content.

No Internal Access

We do not use your formulation data for training, analytics, or any purpose other than delivering your results back to you.

You Always Own Your Data

Export your formulations at any time in full. If you leave, your data leaves with you.

Your formulationPrivate workspaceEncrypted vaultOnly you hold the keyYour account
Private workspace per account
GDPR Compliant
India DPDP Compliant
Hosted on Google Cloud
SOC 2 Type I — Planned

We never share formulation data with ingredient suppliers, distributors, or any third party. Platform neutrality is architectural.

GET STARTED

Start free. Upgrade when you're ready.

Registration is free, no credit card, no commitment. When you're ready to subscribe, we'll walk you through your options personally so you can decide with full information.

Free access

7-day trial · 6 credits · no credit card

Request access in under a minute. Access requests are approved within 24 hours. Your 6 trial credits are live the moment you activate your account. 7-day trial, no credit card.

Subscription access

When you're ready to subscribe, request access. We'll explain the credit system and recommend the right tier for your workflow. No pressure, full information.

Plans from $25/month — see full pricing →

Quick Formula — 1 credit — up to 3 exploration variants at no extra cost

Intelligence Brief — 3 credits — lock a variant to add full safety, regulatory & market depth

Dossier — 5 credits — lock a variant to add stability protocol, manufacturing brief, research citations

Formulation Partner — 1 credit per 5 exchanges

Plans activate automatically once payment is confirmed.

The intelligence layer your R&D team didn't know it was missing.

Request access. Experience the platform. Request a subscription when you're ready.

✓No credit card required
✓7-day trial with 6 credits on activation; then Free at 3 credits/month.
✓Subscribe at any time

*theformulator.ai has no commercial relationships with ingredient suppliers. Ingredient recommendations are ranked by performance data, market frequency, and regulatory compliance, never by commercial agreements. Platform neutrality is architectural.

Frequently asked questions

Is an AI-generated cosmetic formulation regulatory-compliant?

The formulation is *screened*, not certified. Each ingredient's concentration is checked against the restricted-substance annexes and positive lists of the markets you name in the brief — EU Annexes III, V and VI, China's IECIC, Japan's MHLW standards, Korea's MFDS rules, India's IS 4707 — and any breach is reported with the governing entry cited. What it cannot do is replace a Product Information File, a safety assessment under Article 10, or a CPNP notification. Treat the output as a compliant starting point that a qualified safety assessor still signs off, and re-run the screen if you add a market later.

What do I need to put in the brief?

Product format, claims, target markets, sensory goals and technical constraints, entered across five structured layers — for example "sulphate-free · no silicones · EU + India · brightening". The brief is parsed for product type, markets, ingredient constraints and claims, which determines which of the 17 data layers activate. You do not need to name ingredients: the engine assembles a candidate pool from surfactant mildness data, preservative pH compatibility, market co-occurrence and safety profiles, and a second independent pass designs phases, concentrations and processing order. Regulatory conflicts — an EU/China tension, an Ecocert constraint, a prohibited ingredient — surface inline while you are still writing the brief. The reference layer includes 25,000+ supplier formulations and 504K+ formulation units extracted from 69,000+ cosmetic patents.

What is the difference between a Quick Formula, an Intelligence Brief and a Dossier?

Depth of report on the same formula. A Quick Formula (1 credit) is the INCI table with phases, pre-validated concentrations, a per-market regulatory pass/fail and the preservation system. An Intelligence Brief (3 credits) adds per-ingredient safety scores, an ingredient-level regulatory table, market co-occurrence data and preservation rationale. A Dossier (5 credits) adds a stability risk assessment across ICH zones, a manufacturing brief, technical ingredient data and PubMed citations per active where literature exists — structured for your regulatory file and contract-manufacturer handoff. You explore up to three variants first and pay only to lock and enrich the one you choose.

Why is the output INCI-only, with no trade names?

Because the platform has no commercial relationship with any ingredient supplier. Trade names that enter through your brief or a supplier data sheet are rewritten to their INCI names and flagged. Ingredients are ranked by performance data, frequency in 1M+ launched products, stability data and regulatory status — the ranking has no supplier field, so no commercial agreement can influence it. An INCI-only formula is also portable: it can go to any supplier and any contract manufacturer without a locked-in raw-material list.

Does it check ingredient safety, or just legality?

Both, and separately. Legality is the regulatory screen against your named markets. Safety is a six-axis hazard score — carcinogenicity, developmental and reproductive toxicity, sensitisation, systemic toxicity, irritation and endocrine disruption — computed for 11,000+ ingredients from ECHA CLP classifications, REACH SVHC listings, IARC and NTP monographs, and CIR final safety assessments. The two can disagree: an ingredient can be permitted in a market and still carry a hazard flag, or be restricted for a reason that is not toxicological. The report shows both fields so the formulator decides. The scoring methodology is published on this site.