Know how it smells before it exists.

Describe a note in one sentence. Arome designs a new molecule, predicts how it smells and whether it is safe, and maps a route to make it.

Request early access Investors and partners
O

AR-1427 C13H20O

Reading the plume Illustrative

AR-1427, from the brief a long-lasting citrus. Predicted odor, top 6 of 152 descriptors:

  • citrus0.81
  • aldehydic0.62
  • green0.48
  • fresh0.44
  • woody0.33
  • waxy0.21

Share of the plume is the score. Fast strands are volatile. Dense mist is intensity.

Describe it. Arome designs the molecule.

Write a brief in plain words. The generative model proposes a new structure. The predictive models then render how it should smell, how strong it is, how long it lasts, and whether it is safe. Two briefs are ready to try.

Or start from an example

  1. Read the brief
  2. Design a new structure
  3. Predict odor across 152 descriptors
  4. Screen toxicity and biodegradability
  5. Check novelty against PubChem

Illustrative Results on this page are pre-computed examples. The live models run in the Arome platform.

OOO

AR-3071 C12H14O3

Opening, 0 min

Every wisp is a prediction.

Smell has always been invisible until a molecule exists and someone smells it. Arome's predictive models read a structure and return its odor as numbers, before anything is made. The plumes on this page draw those numbers.

Rise speed is volatility. Melon and fruity strands leave first.
Density is intensity. A strong material fills the air.
OO

AR-5520 C12H20O2

Predicted response across a panel of 412 olfactory receptors Illustrative

AR-5520, predicted from its structure alone. Below the words, a receptor-level model predicts how the molecule may activate each of 412 olfactory receptors, the sensors in the nose where smell begins.
Odor profile
Scores across 152 odor descriptors, from marine to musk.
In the plume: each strand's colour and share.
Odor intensity
How strongly the molecule smells.
In the plume: the density of the mist.
Volatility and longevity
How fast it leaves a blotter or skin, and how long it stays.
In the plume: rise speed, and the drydown over time.
Toxicity
Predicted safety flags, before anything is made.
Biodegradability
Predicted breakdown in the environment.
Novelty
Every candidate is checked against PubChem, so what you get is new.

Every plume has a source.

A molecule that cannot be made is only a picture. For every candidate, Arome plans how to make it: by chemistry, step by step, back to materials you can buy, or by biology, through enzymes, from renewable feedstocks.

OO

AR-5520 C12H20O2 · target

Retrosynthesis

A chemical route, planned backwards from the target to starting materials you can buy.

  1. Step 2Radical addition to acrylic acid, then ring closure to the lactone
  2. 2,6-dimethylhept-5-en-1-olC9H18O
    acrylic acidC3H4O2Purchasable
  3. Step 1Reduction of the aldehyde with sodium borohydride
  4. 2,6-dimethylhept-5-enalC9H16OPurchasable
Illustrative Two steps from two purchasable starting materials.

Biosynthesis

An enzymatic pathway in an engineered microbe, traced back to a renewable feedstock.

  1. EnzymeLactonase closes the ring
  2. 4-hydroxy acidC12, open-chain precursor
  3. EnzymeCarboligase adds three carbons, a reductase sets the alcohol
  4. C9 terpenoid aldehydefrom oxidative cleavage
  5. EnzymeTerpene synthase, then a cleavage oxygenase
  6. Geranyl diphosphateC10, mevalonate pathway
  7. HostEngineered yeast turns sugar into acetyl-CoA
  8. Sugar from beet or canefermentation feedstockRenewable
Illustrative Planned from the target down. Made from the feedstock up.

Agents that run the loop, so chemists run the decisions.

Give Arome a brief. Agents generate candidates, predict every property, screen for safety and novelty, plan routes and write the report. Your team reviews a short list, not a pile.

Brief: watermelon and coconut, long-lasting, for a body wash Illustrative run

  1. Generatenew structures for the brief24
  2. Predictodor, intensity, volatility24
  3. Screen safetytoxicity, biodegradability18
  4. Check noveltyagainst PubChem14
  5. Match the briefmelon top, creamy base5
  6. Plan routeschemistry and biology2
  7. ReportAR-5520 recommended, AR-5513 as alternate1
  • AR-5501No short route
  • AR-5502Off-brief
  • AR-5503No short route
  • AR-5504Toxicity flag
  • AR-5505Off-brief
  • AR-5506Off-brief
  • AR-5507In PubChem
  • AR-5508Off-brief
  • AR-5509Off-brief
  • AR-5510Off-brief
  • AR-5511Persistent
  • AR-5512Toxicity flag
  • AR-5513Alternate
  • AR-5514Toxicity flag
  • AR-5515In PubChem
  • AR-5516Toxicity flag
  • AR-5517Off-brief
  • AR-5518Off-brief
  • AR-5519Off-brief
  • AR-5520Recommended
  • AR-5521In PubChem
  • AR-5522No short route
  • AR-5523Persistent
  • AR-5524In PubChem

Agents also take the smaller jobs.

  • Screen a candidate set against your own safety rules.
  • Compare two routes by steps and starting materials.
  • Answer questions about a molecule, such as why it reads green and what to change.
  • Draft a summary of a run for the team.

Who Arome is for

  • Fragrance houses and ingredient makers

    New, ownable materials, from a brief to a route in one loop. Built for R&D leads and fragrance chemists.

    Request early access

  • Perfumers and brands

    New notes to compose with, described in your own words. Arome makes the palette. The perfumer makes the perfume.

    Join early access

  • Investors

    The perfumer's palette has barely grown in decades. We are adding to it, one made-to-be-made molecule at a time.

    Start a conversation

  • Partners

    Ingredient makers, fermentation and biotech teams, and research labs who turn predictions into material.

    Propose a partnership

New molecules for a wider sensory world.

We are building the whole path from a sentence to a molecule that is predicted safe, predicted to smell right, and has a mapped route to be made. We want to build it with people who make things.

Investors

If new materials for fragrance, and the AI that designs them, sit in your field, we would like to talk.

Partners

Ingredient makers to scale materials. Fermentation and biotech teams to run pathways. Research labs to test predictions against the nose.

Start a conversation

Ask for early access, or tell us what you are working on.