Download Profile
Osmo Logo
👃

Osmo

Digital Olfaction

Giving computers a sense of smell. Using Graph Neural Networks to map the relationship between molecular structure and odor perception.

View Research ↗

About This Tool

Osmo is a deep tech company spun out of Google Research. While computers can see (Computer Vision) and hear (Audio AI), they haven’t been able to smell—until now. Osmo creates a “Principal Odor Map,” allowing AI to predict how a chemical molecule will smell without a human ever sniffing it.

How to Use

  1. 1. (For Industry) Define a target scent profile (e.g., “Fresh Rose”).
  2. 2. The AI searches billions of molecular combinations in the Scent Graph.
  3. 3. It predicts the odor descriptors of new, unmapped molecules.
  4. 4. Scientists synthesize the candidate molecules in the lab.
  5. 5. Validate safety and sustainability for use in perfumes or repellents.

Key Features

🗺️ Odor Map
⚗️ Scent Teleportation
🦟 Insect Repellents

Related Tools

🧬

AlphaFold

Protein Structure

🧪

Isomorphic Labs

Drug Discovery

Additional Information

Scroll

Use Cases

Osmo is revolutionizing the fragrance and flavor industries (F&F). It helps companies find safer, non-toxic, and biodegradable alternatives to popular scents that are currently derived from endangered plants or harsh petrochemicals.

Graph Neural Networks

The core technology treats molecular structures as graphs. By training on thousands of known molecule-odor pairs, the AI learns the hidden rules of smell. It can predict that a specific arrangement of atoms will smell like “grassy vanilla” versus “rotten eggs.”

Disease Detection

Beyond perfume, digital olfaction has massive potential in healthcare. Human diseases often emit specific volatile organic compounds (scents). Osmo’s technology could eventually lead to “electronic noses” that detect diseases like cancer or diabetes purely through breath analysis.

Limitations

Smell is subjective. While the AI can predict molecular properties, human perception varies based on genetics and culture. Additionally, synthesizing predicted molecules is still a physical lab process that takes time and resources.

Insect Control

One of Osmo’s most impactful projects is creating better insect repellents. By understanding what scents mosquitoes hate (and why), they can design super-repellents that are safe for humans but highly effective at preventing the spread of malaria and dengue.