News / Standardization

From standard to ecosystem: our report to the ISO/TC 46 plenary in Milan

At the ISO/TC 46 plenary week in Milan we reported on a year of building around ISO 24138:2024: implementations for six language targets, similarity search at scale, the draft Discovery Protocol and first pilots in publishing, science and open content.

From 6 to 10 July 2026 the Italian standards body UNI hosted the plenary week of ISO/TC 46, the technical committee for information and documentation, in Milan. TC 46 is the committee behind ISO 24138:2024, the International Standard Content Code (ISCC). Delegates from national standards bodies, libraries, archives and publishers met in person and online, and much of the week's discussion turned on artificial intelligence and how the committee's standards should respond to it. We reported on ISCC under the title "A year of progress": what has been built around the standard since the committee last met.

One year, measured

  • More than 2,800 commits across the ISCC codebase
  • More than 20 active open source repositories
  • Three new building blocks: iscc-lib, iscc-hub and iscc-search
  • Six language targets for the standard in iscc-lib

One specification, six language targets

The Python reference implementation, iscc-core, reached version 1.3.0 with full test coverage and remains the reference for ISO 24138:2024. The Rust core, iscc-lib, reached version 0.5.0. It is still experimental and its interfaces may change, but it is conformance-tested against the reference implementation and brings a high-performance core to the browser through WebAssembly and to Go and Swift through native bindings. Whether a file is processed in a browser, on a phone or in a data centre, the same input gives the same code.

Content-first metadata discovery

Resolution starts from an identifier. Discovery starts from the content. The draft ISCC Discovery Protocol specification describes how to publicly declare ISCC-CODEs to interconnect existing and authoritative content registries across sectors.

  1. A declaring party signs an IsccNote with their own key. An IsccNote contains extended 256-bit ISCC-UNITs and an optional link for discovering metadata and services in established registries.
  2. An ISCC-HUB records the IsccNote in a public, auditable transparency log and issues an ISCC-ID to the declarer.
  3. Anyone can resolve metadata and services in established registries using the ISCC Discovery Protocol.

Three ways to discover

  • Identifier lookup

    Resolve an ISCC-ID straight to its declaration and associated registries.

  • Similarity discovery

    Match an ISCC-CODE to related content. Designed to find different versions of the same content.

  • Instance-Code query

    Find declarations of identical files by different parties based on their cryptographic hash.

Built on three layers

  • HUBDeclaration, timestamping, and transparency log
  • GATEWAYRouting and resolution to established infrastructure
  • REGISTRYEstablished infrastructure with authoritative metadata

Generating a code stays local and needs no registration; declaring one to obtain an ISCC-ID is an optional, separate step. The Discovery Protocol itself is a draft, and developed in the open as a set of ISCC Enhancement Proposals.

Similarity search at web scale

A hub or a registry has to answer a similarity query fast, at web scale. ISCC Search is a new open source engine for scalable ISCC matching, built to hold billions of ISCC-UNITs and return near-duplicates in milliseconds on a single machine. It is in beta but its interfaces are fairly stable.

See how ISCC matching works by dropping in two files and comparing them at web.iscc.io.

Matching meaning, experimentally

ISO 24138:2024 reserves a Semantic-Code but does not standardize it yet. Our work here is experimental. Our iscc-sct project implements a Semantic-Code for text to match the same work across paraphrases, rewrites and even languages. iscc-sci, a Semantic-Code for images, is designed to recognize the same picture through crops, filters and stylistic variation. Both are proofs of concept, documented so the community can evaluate them before any standardization is considered.

Put to work

Three sectors are putting ISCC to work outside the lab. In scholarly publishing, a pilot with a large scientific publisher applies ISCC to research papers. The findings have been shared with the European Commission. In science and genomics, ISCC-SUM computes the Data-Code and Instance-Code of large files in a single pass and is integrated into the Galaxy platform. The BIO-CODES initiative works with EMBL and Euro-BioImaging on identifiers for research data. For open and cultural content, the CommonsDB registry is piloting ISCC with a corpus of over 5 million assets.

With the standards community

ISCC has to work alongside other standards and initiatives, so we build in the open with the organizations behind them.

  • ISCC is the first soft binding registered with the Coalition for Content Provenance and Authenticity (C2PA), so a content credential can be looked up from the content itself when embedded metadata has been stripped.
  • The Discovery Protocol builds on W3C Decentralized Identifiers and Controlled Identifiers.
  • Our talk at the FDO Forum 2026 traced the path from a content code to a FAIR Digital Object.
  • New enhancement proposals are being built to feed the future standard's track at ISO.

What comes next

In one year the published standard has gained: implementations for six language targets, search that scales, a draft protocol for discovery and first pilots in publishing, science and open content. The committee meets again in May 2027 near Paris, and we intend to bring the next year's results with us.

Everything we presented is open source at github.com/iscc. If you work with digital content and want to put the standard to work, or help shape the enhancement proposals, get involved.