Inside Osmo’s Groundbreaking Scent Teleportation Technology

Earlier this year, Osmo introduced the concept of Scent Teleportation, a technology designed to capture a scent in one place and diffuse it in another. This groundbreaking project has sparked lively discussions and curiosity around how it works and how close Osmo is to achieving its goals. Now, Osmo is offering a closer look at the complex technology behind Scent Teleportation, sharing recent milestones and revealing plans for public demonstrations.

How Scent Teleportation Works: The Process Explained

Osmo has successfully achieved scent teleportation in its labs, starting with a simple slice of coconut as the first transmitted scent from one side of the lab to the other. Although the idea of “sending a scent” sounds straightforward, the actual process involves a sophisticated combination of AI, chemical analysis, and precise molecular reconstruction.

Capturing the Scent

The process begins by selecting a scent and introducing it to a device called the GCMS (Gas Chromatography-Mass Spectrometry). If the sample is a liquid, it’s directly injected into the machine. For solid objects like fruits, headspace analysis is employed: a small, sealed chamber collects and concentrates scent molecules from the air surrounding the object. These molecules are then absorbed into a tube, capturing the complex aroma profile of the sample.

Breaking Down the Scent to Molecular Data

The GCMS device analyzes the sample, separating the scent into its constituent molecules and identifying them based on molecular weight and structure. This analysis produces a unique "molecular fingerprint," a detailed readout of each compound in the scent. This data is then uploaded to Osmo’s cloud storage, forming a digital dataset that represents the scent on a molecular level.

Mapping Scent Data with AI

The molecular fingerprint is interpreted and mapped onto Osmo’s Principal Odor Map, an AI-powered tool that plots scents within a multidimensional "scent space." The Principal Odor Map utilizes machine learning algorithms trained on thousands of scents to predict how a specific combination of molecules will smell. It categorizes the scent into recognizable odor families (such as fruity, floral, or woody) and identifies subtle nuances that are often challenging to capture. This step is crucial for ensuring the digital scent profile matches the original aroma as closely as possible.

Recreating the Scent

Once the scent profile has been mapped, it’s sent as a "recipe" to a Formulation Robot, which is responsible for reconstructing the scent. The robot combines a precise mix of base scent compounds, using the recipe to recreate the original scent’s molecular structure. The Formulation Robot’s capabilities include adjusting concentration levels and balancing complex notes, ensuring that the recreated scent has the same intensity and character as the original.

Fine-Tuning Through AI and Machine Learning Deformulation
Osmo uses machine learning deformulation models to refine the scent reconstruction process. These models predict the best formulation based on the GCMS data and continuously learn from each trial, becoming better at matching subtle scent profiles over time. By refining the process in every trial, Osmo’s AI-driven approach gradually reduces the gap between the original scent and the recreated version, achieving higher fidelity with each iteration.

Testing and Comparison for Authenticity
Finally, Osmo’s team conducts a comparison between the original sample and the recreated scent. This is done by human testers who evaluate the recreated scent's accuracy, focusing on odor family alignment, intensity, and the presence of subtle scent notes. Osmo’s goal is to replicate every nuance, including difficult-to-capture compounds like sulphurs in tropical fruits. Each trial provides new data to enhance their understanding of molecular structures and further refine the scent recreation process.

Building the World’s Largest Scent Data Bank

Over time, Osmo’s collection of molecular data has grown into the world’s largest AI-compatible scent database. This expansive data bank is used to continuously train Osmo’s algorithms, allowing them to predict and recreate scents with increasing precision. This iterative process not only strengthens Osmo’s ability to capture nuanced aromas but also reduces reliance on human intervention, moving toward a fully automated scent recreation process.

Overcoming Technical Challenges

Despite progress, Osmo’s scent digitization journey has not been without its challenges. Achieving precise recreations requires addressing several technical hurdles:

• Untraceable Molecules: Some molecules that play critical roles in scents are hard to detect and reproduce accurately. For example, sulphurs in tropical fruit are notoriously hard to detect and reproduce.

• Undiscovered Molecules: Certain molecules in everyday scents remain unidentified, and Osmo is continuously working to reduce this number through extensive data collection and analysis.

To tackle these challenges, Osmo has developed the largest AI-compatible scent database, which powers machine learning models for decoding and recreating scents. By using Machine Learning Deformulation models and advanced algorithms, Osmo is refining its ability to “translate” molecular data into accurate scent recreations.

The Principal Odor Map and Monthly Scent Trials

The Principal Odor Map plays a central role in Osmo’s process. This multidimensional map allows Osmo to understand and predict scent profiles more accurately. Monthly Scent Teleportation trials put the technology to the test, providing insights that help Osmo refine its algorithms and make the system faster and more automated.

Public Demos: Experience Scent Teleportation First-Hand

Excitingly, Osmo plans to open select demos to the public soon. Participants will have the opportunity to choose a scent—such as a flower or fruit—and observe as Osmo’s technology captures and recreates it in real-time. Participants will be asked to evaluate the accuracy of the reproduced scent, offering valuable feedback that will drive further improvements.

Rapid Progress and Future Potential

In just six months, Osmo has made significant advances in Scent Teleportation. Each month, the system becomes more efficient, automated, and capable of handling complex scents. Osmo envisions a future where sharing a scent is as simple as sending a photo or a song—an innovation that could redefine the way we connect through sensory experiences.

Looking Ahead

This research is crucial because scent plays a powerful role in memory, emotion, and sensory experience, yet it remains one of the least explored senses in digital technology. By enabling scent transmission, Osmo’s Scent Teleportation could open new possibilities in virtual communication, healthcare, and immersive experiences, allowing people to share and experience scents as easily as they would a photo or sound.

Osmo’s work on Scent Teleportation demonstrates a pioneering step in olfactory technology, bringing us closer to a world where scents can be shared across distances. With ongoing improvements and public testing on the horizon, Osmo is steadily transforming an ambitious concept into a tangible reality, promising a new way for people to connect through scent.