---
id: engines/clay
about:
  label: Clay
  description: >-
    Open Earth observation foundation model, now a program of Renaissance
    Philanthropy: it turns satellite or aerial imagery with any number of bands,
    along with its place and date, into 1,024-number embeddings.
  tags:
    - satellite
    - foundation model
    - embeddings
    - sentinel
    - land cover
tool:
  form: open model
  group: Earth observation and AI
  homepage: 'https://clay-foundation.github.io/model/'
  licence: >-
    Apache-2.0 (code and weights); CC-BY-4.0 (documentation); ODC-BY (v0
    embeddings); CC-BY (v1.5 embedding sets)
  cost: free
domains:
  - places
  - ecology
open:
  source: open
  data: open
  api: unknown
  participation: open
  licence: Apache-2.0; embeddings ODC-BY (v0) or CC-BY (v1.5)
  note: >-
    Source Cooperative holds precomputed embeddings for the v0 training data,
    for the whole NAIP archive (2010 to 2021) from v1.5, and for Sentinel-2
    worldwide, one per month from January 2017 to April 2026, from v1.5 via
    lgnd.ai; for other imagery you run the model yourself.
relations:
  - predicate: alternative_to
    target: engines/tessera
  - predicate: alternative_to
    target: engines/alphaearth-prithvi
provenance:
  origin: agent
  sources:
    - title: Clay Foundation Model (documentation home)
      url: 'https://clay-foundation.github.io/model/'
    - title: Clay-foundation/model (README)
      url: 'https://github.com/Clay-foundation/model'
    - title: Model release notes (Clay v1.5 specification)
      url: 'https://clay-foundation.github.io/model/release-notes/specification.html'
    - title: Training Data
      url: 'https://clay-foundation.github.io/model/release-notes/data_sampling.html'
    - title: Explore embedding space of CLAY Encoder for different sensors
      url: 'https://clay-foundation.github.io/model/tutorials/embeddings.html'
    - title: made-with-clay/Clay (Hugging Face model card)
      url: 'https://huggingface.co/made-with-clay/Clay'
    - title: Clay Model v0 Embeddings (Source Cooperative)
      url: 'https://source.coop/clay/clay-model-v0-embeddings'
    - title: NAIP data embedded with Clay v1.5 (rev2) (Source Cooperative)
      url: 'https://source.coop/clay/clay-v1-5-naip-2'
    - title: LGND Embeddings (Source Cooperative)
      url: 'https://source.coop/clay/lgnd-embeddings'
  created: '2026-10-21'
  reviewed: '2026-10-21'
  reviewed_by: 'Claude, from the sources listed'
  status: draft
---
Clay reads a square of satellite or aerial imagery together with where and when it was taken, and returns an embedding: a list of 1,024 numbers that summarises what is on the ground in that square at that moment.[^1,8] Squares that look alike to the model get similar numbers, so the embeddings can be searched by similarity or fed to small classifiers.[^1,9] The model is a vision transformer, the same family of network used for photographs, trained without labels by hiding three quarters of each image and learning to fill it back in.[^1,3]

Versions v0 to v1.5 were a fiscally sponsored project of the non-profit Radiant Earth; Clay is now a program of Renaissance Philanthropy.[^1] The current release, v1.5, was trained in September 2024 on 70 million image chips chosen to cover the world's land cover types.[^3,4] Its sources were Sentinel-2, Sentinel-1 radar, Landsat 8 and 9 and MODIS worldwide, plus NAIP aerial photos of the USA and LINZ aerial photos of New Zealand, with ground detail from under half a metre to 500 metres per pixel.[^4] Instead of a fixed sensor, the model is told the wavelength of each band, so it accepts any number of bands from many kinds of sensor.[^3] The documentation credits Prithvi among the models whose design it drew on.[^3,4]

The code installs as a Python package from GitHub and the v1.5 weights download from Hugging Face; the encoder alone is 1.25 GB.[^1,3] It has three main uses:[^1]

- finding features such as surface mines, aquaculture or feedlots by similarity[^1]
- training a small model on the embeddings to classify crop type or land cover, estimate biomass, or detect change such as deforestation, wildfire, flooding or new building[^1]
- using Clay as the backbone of a larger model[^1,6]

Each embedding describes one place at one time, so following a place through the seasons means comparing embeddings from several dates.[^3] Some are already computed and published on Source Cooperative.[^7,8,9] Clay used v1.5 to embed the whole NAIP aerial archive of the USA from 2010 to 2021.[^8] The company lgnd.ai publishes monthly v1.5 embeddings of Sentinel-2 worldwide from January 2017 to April 2026, under CC BY 4.0.[^9] Embeddings for all the training data of the older v0 model are there too, under ODC-BY.[^1,7]

Code and weights are under Apache 2.0 and the documentation under CC-BY 4.0.[^1,2] The training data covered only land and coastal water, left out the poles, open ocean and night-time scenes, did not set out to include extreme events, and had at most six dates per location.[^3]