About us

The Open Databases Integration for Materials Design (OPTIMADE) is a consortium of materials database providers with the aim of making materials databases interoperable.

The OPTIMADE API is a standard API that unifies searching and retrieving information from materials databases. → More about the motivations behind the OPTIMADE API

The latest release of the specification is v1.3.0 (December 2025).

0	0	0
PROVIDERS	DATABASES	STRUCTURES

Status as of Sep 2026, taken from the OPTIMADE providers dashboard.

Why it is worth it

If you consume data, one filter query works everywhere. You write it once and send it to every database that implements OPTIMADE, instead of learning a different API, a different query syntax and a different data model for each one. Tools already exist that query every known OPTIMADE database at once and hand you back a single set of results.

If you provide data, implementing OPTIMADE makes your database visible to everyone already using those tools, without you building or maintaining a client of your own. You keep your own API, your own data model and your own terms of use; OPTIMADE is an additional, standard way in. The reference implementation does most of the work, and once you are registered every OPTIMADE client can reach you.

Try it now

This asks the Crystallography Open Database for binary silicon oxides, and is copy-pasteable as it stands:

$ curl -G 'https://www.crystallography.net/cod/optimade/v1/structures' \
    --data-urlencode 'filter=elements HAS ALL "Si","O" AND nelements=2' \
    --data-urlencode 'response_fields=chemical_formula_descriptive,nelements' \
    --data-urlencode 'page_limit=1'

{
  "data": [
    {
      "id": "1010921",
      "type": "structures",
      "attributes": {
        "chemical_formula_descriptive": "O2Si",
        "nelements": 2
      }
    }
  ],
  "meta": {
    "data_returned": 394,
    "data_available": 535065
  }
}

Same query, any other provider: change the base URL. data_returned is how many entries matched, data_available how many the database holds.

The structures themselves come back the same way. Ask for response_fields=lattice_vectors,cartesian_site_positions,species_at_sites,species and you get the unit cell, the Cartesian coordinates of every site, and what sits on each one — species_at_sites names entries in species, which is where the chemical symbols are. Those fields mean the same thing at every provider, which is the whole point of the standard.

Response retrieved on 11 September 2026 and shortened for clarity — the real one also carries links and relationships for each entry, and the counts change as the database grows.

Query everything at once

To search across the whole federation rather than one database at a time:

• the providers dashboard lists every known provider, what each one serves and which version it speaks;
• optimade.science queries all known databases simultaneously from your browser;
• the optimade-python-tools client does the same from Python or the command line.

All of these are compared on the clients page.

Where to go next

• Query data → clients
• Serve your data → resources
• Get involved → community

If you use OPTIMADE in published work, please see how to cite.

Motivation

Designing new materials suitable for specific applications is a long, complex, and costly process. Researchers think of new ideas based on intuition and experience. Their synthesis and evaluation require a tremendous amount of trial and error. In the last few years, there has been a major game change in materials design. Thanks to the exponential growth of computer power and the development of robust first-principles electronic structure codes, it has become possible to perform large sets of calculations automatically. This is the burgeoning area of high-throughput ab initio computation. Such calculations have been used to create large databases containing the calculated properties of existing and hypothetical materials, many of which have appeared online.

Those databases grew up independently, each with its own API, query language and data model, which makes using more than one of them at a time far harder than it should be. OPTIMADE exists to remove that barrier.

Get involved

Everyone is welcome. We meet online roughly once a month, and we would like to help you create and register your own OPTIMADE API implementation.

The next meeting will take place on Friday October 2^nd 2026 at 14:30 UTC/16:30 CEST on Jitsi

The community page says which channel to use for what, and how to ask for access to our Slack workspace. The most recent OPTIMADE CECAM workshop was held on the campus of the University Grenoble Alpes, June 29 – July 3 2026; all of our workshops are listed on the events page.