ScentHash

Experience the future of scent

Encode any scent into a digital recipe — shareable, comparable, reproducible.

ScentHashScentHash
ScentHash · protocol v0.1

Scent, encoded.

A digital representation layer for scent — built to capture, identify, compare, share, and connect smell with software and compatible devices.

REAL SCENT→ SCENTHASH→ DIGITAL ID→ API→ DEVICE
image
JPEG
sound
MP3
video
MP4
scent
ScentHash
Resolved entry
Sandalwood
IDSH-0001
HASH—
Evaporation profile topheartbase
spray30mhours
Live encoder · v0.1

Turn a smell into data — right here.

Type any scent in words. It's encoded into a base-note recipe, hashed with SHA-256, and given a registered ID.

The protocol

A representation layer for scent.

From a real-world scent to a digital identity — and, one day, back out to hardware. Tap any stage.

Why a digital layer

Everything else has a format. Scent doesn't.

Images
JPEG · PNG · SVG
Audio
MP3 · WAV · FLAC
Video
MP4 · MOV · WEBM
Documents
PDF · DOCX · TXT
Scent
— none —

ScentHash proposes a common digital representation for scent — a way to capture, compare, and exchange it as data.

How it works

From a scent to a code, in three steps.

01 / compose

Compose your scent

Mix from the forty base notes across top, heart, and base layers until it totals 100%.

02 / register

Register it

The scent gets a registered ID, a SHA-256 fingerprint, and a timestamp — a traceable record.

03 / play

Find it & play it

It appears in your library, gets rated by the community, and — when a device exists — plays anywhere.

Digital identity

How a scent becomes a digital identity.

Encode any scent and it resolves to a structured record with a stable identity — a representation, a fingerprint, and a version. Try it:

A cryptographic hash like SHA-256 fingerprints the digital representation — not the physical smell.

What it enables

What a digital representation makes possible.

Once a scent is structured data, it inherits everything data can do.

What's real today

Built, and proposed — kept separate.

Live now
  • Compose — build a scent from base notes
  • Register — save it with an ID
  • Hash — SHA-256 of its digital recipe
  • Library — browse registered scents
  • Similarity — compare by fingerprint
Proposed
  • Standard — the layered representation
  • API — a public, documented endpoint
  • Device — the concept diffuser
  • Hardware — sensing & the physical layer
Who it's for

Three ways into scent technology.

One ecosystem, three paths — experience it, build on it, or engineer the hardware behind it.

Published works

Top community scents.

The library

Ready-made scents.

Each a recipe from the forty base notes. The collection grows as makers register their own.

Maker panel

Compose a scent.

Evaporation profile
Total0%
Similarity engine

Find the nearest scent.

Every scent is a fingerprint across forty base notes. Pick one; the library ranks by how close each one smells.

Nearest scents

Base palette

Forty base notes.

The primary notes every composition references. The standard's foundation.

The proposal

ScentHash is a representation layer, not a database.

A scent is described in layers. Today only the first exists in full; the deeper layers are proposed and grow as chemistry and devices are mapped in. This is an experimental draft, not a ratified standard.

The 40-note model is an experimental v0.1 representation layer — not a claim that every smell can be reduced to 40 notes.

Proposed interface

The API for scent.

A proposed surface so any app, video, or device can speak ScentHash. Documentation only — no live endpoint yet.

Specification

ScentHash Specification v0.1 (draft).


    

  

  

    
Capture
Describe a smell. Get a ScentHash.
Write a scent in words; it's encoded into a base-note recipe, hashed with SHA-256, and given an ID. A live AI model does the mapping, deterministically.

    
    

    
    

  




  

    
Your work
My scents.
Everything you've registered — public or sealed — lives here with its code and fingerprint.

    

  


  

    

      
For everyone

      
Scent, as an experience.

      
No chemistry, no code. Browse ready-made scents, see how they relate, and give any smell a name you can keep.

    

    

      
      
      
    

    

  


  

    

      
For developers

      
Build with scent data.

      
ScentHash turns a scent into a compact, comparable record — a base-note recipe, a fingerprint vector, and a content hash. The encoder works today; a public API and SDKs are in development.

      Platform · in development
      

        
        
      

    


    

      
The data model

      
Every scent resolves to one record. This is the shape the encoder returns today.

      
{
  "scentHash": "SH-000123",
  "version": "0.1",
  "layer1": { "semantic": ["rose", "sweet"] },
  "layer2": {
    "recipe": [
      { "code": "BN-05", "pct": 60 },
      { "code": "BN-17", "pct": 40 }
    ]
  },
  "fingerprint": [ /* 40 values, 0-100 */ ],
  "hash": "sha256..."
}

    


    

      
Encoding interface

      
Text in, structured scent out. Reference shape — the public endpoint is in development.

      
POST /encode
Content-Type: application/json

{ "input": "fresh rain on warm earth" }

  ->  { scentHash, layer1, layer2, fingerprint, hash }

      
No public keys or rate limits yet. Early-access partners can request access.

    


    

      
What's live · what's coming

      

        

          
Live now

          
    
            
  • AI text-to-scent encoder (deterministic) — recipe, fingerprint, hash
    • 
            
  • The forty-base-note standard (v0.1)
    • 
            
  • Registered library of community scents
    • 
          

        

        

          
In development

          
    
            
  • Public REST API and authentication keys
    • 
            
  • Client SDKs (JavaScript, Python)
    • 
            
  • Rate limits, webhooks, versioning
    • 
            
  • Decode — record back into a device recipe
    • 
          

        

      

    


    

      
Explore the building blocks

      
The standard and the base palette are the foundation everything encodes against.

      

        
        
      

    

    

  


  

    

      
For hardware

      
The physical layer of digital scent.

      
A ScentHash record is a recipe across forty base notes. A device holds those notes and doses them to reproduce a scent. The dosing is computed today; the device itself is a prototype.

      Prototype · research
      

        
        
      

    


    

      
From record to scent

      
How a digital record becomes something you can smell.

      

        
01
Record
A scent as a base-note recipe and fingerprint.

        
→

        
02
Dose
The recipe becomes measured amounts, note by note.

        
→

        
03
Diffuse
The device releases the blend — the scent, reproduced.

      

    


    

      
Where the hardware stands

      
An honest map of what exists, what's prototyped, and what's still research.

      

        
Current
The standard & dosing
The forty-note recipe and its dose balance are computed live in the app today.

        
Prototype
The concept diffuser
A device holding the forty notes, mixing doses from a recipe — simulated in-app, not yet built.

        
Research
Scent sensing
Capturing a real smell into data with sensor arrays — an open research problem.

        
Future
Consumer playback
Pairing a compatible device and playing any scent anywhere.

      

    


    

      
For labs, universities & device makers

      
ScentHash is a platform to build the physical layer on — not a store. We're looking for research and hardware partners.

      

        
        
      

    

    

  


  

    
Devices
Play a scent on a device.
A device holds the forty base notes and mixes measured doses from a recipe — the balance below. Hardware pairing arrives when a compatible device ships.

    

      

        coming soon
        

        
Pair a diffuser

        
Scan a device key to link a scent diffuser. Available once a compatible device is released.

      

      

        
The balance

        
Pick a scent to see its doses.

        
        

        
        

      

    

    

      
Device Prototype
Concept Prototype

      
A hardware platform for scent.

      
The phone and web are the control interface. Scent is produced by a dedicated device that holds replaceable note cartridges, doses them by the recipe, and diffuses the blend. This is a concept prototype.

      

        ScentHash Platform→
        Mobile / Web→
        ScentHash Device→
        Scent Output
      

      
The interface controls; the device delivers. The phone never produces scent itself.

      

        

          
Modular cartridges

          
Each scent family is an independent, replaceable cartridge. The architecture scales to more notes over time.

          
CitrusWoodyFloralFreshCustom+

        

        

          
For hardware manufacturers

          
Build the device to be API-controlled and future-ready. Target attributes:

          
ModularScalableManufacturableMaintainableConsumer-safeEnergy-efficientBLEUSB-CFirmware-upgradableAuto-shutoff

        

      

      
Proposed
As AI systems begin connecting to the physical world through open hardware standards, scent still has no common digital representation. ScentHash is designed to be that interoperability layer — so future scent sensors, AI agents, and applications can communicate through one shared scent representation.

      

        
Build the ScentHash hardware

        
We are looking for hardware partners to turn the ScentHash concept into a scalable, connected scent-technology platform.

        
      

      
Concept prototype for illustration. It does not itself produce scent; a real scent-delivery device requires compatible hardware.

    



  

  

    
  

  

    
  

  

    

      LAB ONLINE
      SIMULATED SENSOR DATA
      MODEL v0.1
    

    
Research prototype · No physical sensor connected

    

      
Experimental lab
Digital olfactory experimentation.

      EXP-0000 · SIMULATED
    

    

      
      
    

    

      Simulation Mode
      Sensor responses are computationally generated from the ScentHash representation. No physical olfactory sensor is connected.
    

    

      
Virtual sensor array · 32simulated

      
System status

      
Real-time signalsimulated

      
Representation cloud · 3Dsimulated

      
40-dimensional fingerprint

      
ScentHash ID

      
Processing pipeline

        

          Raw signal1,200 pts→
          Normalizez-score→
          FeaturesPCA→
          40Drepresentation→
          EncodeScentHash→
          IDSHA-256
        

      

      
Similarity engine

      
Experiment log

    

  








Menu