First, a little humor to set the stage.

NFTs have revolutionized the digital art world. By linking a digital artifact to an onchain identifier, we have unlocked the ability to track its provenance, prove its ownership, and use cryptographic signatures (i.e. attestations) to connect all the offchain lore to them.

But what about collections of art?

Sadly, these have not yet moved onchain.

A collection of NFTs should be a first-class citizen in the web3 world, yet NFT collections live as centralized data in web2 databases, far off the chain.

Atomic Form wants to change this. Today, we are introducing a new primitive to bring NFT collections onchain.

Unlike an NFT, which is an individual token on a blockchain, a collection is a contextual grouping of NFTs. Collections list multiple NFTs along with their metadata, such as title, description, or cover image.

However, NFT collections have no on-chain representation yet. Artists, collectors, and curators must choose one of the NFT marketplace platforms and create a collection there. To see an artist’s portfolio or a collector’s or curator’s collection, you must visit a specific NFT marketplace or maybe several of them for each blockchain.

Marketplaces use collections primarily to present their offerings. Their collections live in one of their warehouse databases and share all the disadvantages of Web2 data: they are centralized, ephemeral, and unverifiable. While these drawbacks may be overlooked for the purpose of trading NFT art, in the art world, a collection is much more than a short-lived, volatile data set—it can represent an artist’s lifetime portfolio, tell a collector’s story, or prove a curator’s exhibition history.

Given the higher demands of art world collections, we are long overdue for a fundamental shift in how NFT collections are managed. The current system has six severe limitations that we need to overcome to make collections universally useful: fragmentation, centralization, incompleteness, fragility, inconsistency, and the lack of a canonical identifier.

• Fragmentation occurs because each marketplace has its own collection system, leading to disconnected and proprietary formats. Collections are disconnected from related collections on other marketplaces, making it difficult to track which NFTs belong to a given collection or which collections a given NFT belongs to. This fragmentation limits the growth and richness of data relationships between collections.

• Centralization results from storing collections in marketplace-specific databases, risking data loss if platforms fail. This creates barriers between marketplaces, making it challenging to transfer or manage collections across multiple platforms due to the lack of a unified, portable data model.

• Incompleteness arises because current NFT collections often lack comprehensive metadata and contextual information, hindering the full representation of an artist’s work, a collector’s holdings, or a curator’s exhibition.

• Fragility is an almost natural property of web2 data. A database record can be deleted in an instant, and without a complete and persistent system of logs, it is next to impossible to verify—or even know—that the record ever existed.

• Inconsistency results from different platforms and databases recording and presenting information variably, leading to discrepancies that undermine the integrity of the collection and erode trust among stakeholders.

• The lack of a canonical identifier prevents referring to a collection as a whole. It cannot have an on-chain representation, nor is there a meaningful way of adding data or metadata that belongs to the whole collection. Additionally, without a standard identifier, there's no single source of truth for multiple digital copies of a collection.

Clearly, the current system is not a good foundation for building and maintaining long-lasting collections. Artists, collectors, and curators need a more reliable solution.

Artists frequently see their portfolio scattered across marketplaces. How should they present their work as a single collection? They have no choice but to manually assemble a “meta-collection” to catalog the various marketplace-specific collections. For example, the Brinkman NFT collection exists on 57 marketplaces, including SuperRare, NiftyGateway, Manifold, Rarible, and many more. The Brinkman NFT Catalog as a whole only exists as a table on Google Sheets, manually collected from all the marketplaces.But even when this tedious, manual work is done, and everything is gathered in one place, the collection is still not verifiable as a whole. While it is possible to verify each NFT in a collection individually, the exact contents of a collection remain the artist's unverified intention. Moreover, artists must frequently take action to keep their portfolio complete. Whenever a marketplace disappears, artists who have a collection listed there have to replicate it in another marketplace.

Collectors and curators have no choice but to trust the marketplace platforms. Unfortunately, the web3 idea of zero-trust transactions has yet to be applied to collections. The nature of web2 data makes verifying a collection’s authenticity, ownership, and completeness virtually impossible. The distribution of collections across marketplaces makes verifying any of these attributes even more difficult.

The only conclusion to be drawn from this flawed state of collections is this:

Collections must move onto the chain.

A canonical, on-chain definition of a collection would solve all of the existing deficiencies once and for all. Based on an on-chain collection, a universal collection standard could be established across marketplaces and chains.

Such a standardized, decentralized collection system would also enable essential features and functionalities of collections that the current system does not support. For example, at present, a tweet about an entire collection must be applied to all NFTs in a collection individually. With a decentralized collection, the tweet could be applied in one step.

To arrive at a world of verifiable collections, the NFT art world must find and agree on a universal collection format that covers all the needs of decentralized, verifiable collections.

For this purpose, Atomic Form proposes a system of attestations with a structured format to build a verifiable, immutable record of NFT collections.

The technical details of the proposal are beyond the scope of this article. In a nutshell, we propose a standard attestation and metadata structure for collections that works across chains and can be easily adopted by marketplaces via API providers. The core of an on-chain collection will be a collection ID that includes the chain, smart contract, and token ID to uniquely identify the collection. The structure is designed to cover different use cases, ranging from simple grouping of NFTs to more complex curation and management of collections.

The Brinkman Catalog example represents a simple use case: All NFTs exist on a single blockchain (Ethereum) but are spread across multiple marketplaces. This scenario presents no particular technical challenges, only organizational ones. All NFTs share the same structure, and the primary task here is to consolidate all the information from different marketplaces.

Transitioning single-chain NFT collections on-chain is, therefore, quite straightforward.

A much more complex use case is the Disobedient collection on OpenSea, with NFTs spread across five different blockchains. This scenario is quite involved on the technical level, as the collection must handle five different blockchain systems simultaneously, each with its own NFT standards. For example, unlike Ethereum, Solana doesn’t require a dedicated smart contract to mint NFTs; its built-in Token Program handles this.

Our on-chain collection proposal addresses these diverse requirements, offering the flexibility and scalability needed to manage multichain collections.

Several technical aspects have been considered when crafting the proposed structure, including:

• Handling unfinished collections: Collection owners must be able to add NFTs to a collection.

• Versioning and updates: All changes to a collection, such as transferring ownership, adding NFTs, or adding off-chain artifacts, should build up a traceable and verifiable history.

• Cross-chain compatibility: A collection should be able to contain NFTs from multiple chains.

• IPFS integration: Metadata must live in a content-addressable space to remain retrievable at any time.

These aspects have shaped the design, resulting in a universal and highly flexible solution.

What do NFT artists, collectors, and curators gain through this proposal?

First, collections become transparent and verifiable. By standardizing the collection structure and rooting it on a chain, everyone can inspect the contents and metadata of a collection.

Equally important is decentralization. Collections will no longer be at the mercy of a single company. Decentralized collections will be immune to marketplace shutdowns, securely living on countless copies of the chain and a decentralized storage system.

A standardized collection system also simplifies collection management. Instead of navigating the nuances of dozens of marketplace platforms, we will have a unified way to add NFTs to a collection, update lore, or transfer ownership —transparently and verifiably.

Finally, universal, chain-based collections provide opportunities for enhanced functionality that are not possible with the current system of isolated, marketplace-owned collections. Features like verifications, allow lists, airdrops, mass attestation via collection reference, or usage by AI agents become feasible, none of which work with off-chain, web2-based collections.

For example, an artist might decide to relocate a collection to another chain. Now, two copies of the collection exist. The same happens if a forger copies a collection. Being able to attest which instance is the correct one is crucial. Without a verifiable attestation, humans, algorithms, and AIs would have to rely on the artist’s website, tweets, or other unreliable web2 sources to tell the legitimate instance from the one that’s been forged or obsoleted.

Our proposal opens up a world of possibilities for artists, collectors, and curators. Collections will become much more than an entry in a proprietary database. On-chain collections fundamentally change the way we can work with collections or any verifiable grouping of NFTs.

By addressing the current limitations of NFT collections, we aim to enhance the creative possibilities of NFTs for artists, collectors, and curators alike. We envision a future where NFT curation is more accessible, verifiable, and interoperable across platforms and chains.

So let’s move collections where they belong: on-chain, with the same level of immutability and verifiability as the NFTs they contain.

First, a little humor to set the stage.

NFTs have revolutionized the digital art world. By linking a digital artifact to an onchain identifier, we have unlocked the ability to track its provenance, prove its ownership, and use cryptographic signatures (i.e. attestations) to connect all the offchain lore to them.

But what about collections of art?

Sadly, these have not yet moved onchain.

A collection of NFTs should be a first-class citizen in the web3 world, yet NFT collections live as centralized data in web2 databases, far off the chain.

Atomic Form wants to change this. Today, we are introducing a new primitive to bring NFT collections onchain.

Unlike an NFT, which is an individual token on a blockchain, a collection is a contextual grouping of NFTs. Collections list multiple NFTs along with their metadata, such as title, description, or cover image.

However, NFT collections have no on-chain representation yet. Artists, collectors, and curators must choose one of the NFT marketplace platforms and create a collection there. To see an artist’s portfolio or a collector’s or curator’s collection, you must visit a specific NFT marketplace or maybe several of them for each blockchain.

Marketplaces use collections primarily to present their offerings. Their collections live in one of their warehouse databases and share all the disadvantages of Web2 data: they are centralized, ephemeral, and unverifiable. While these drawbacks may be overlooked for the purpose of trading NFT art, in the art world, a collection is much more than a short-lived, volatile data set—it can represent an artist’s lifetime portfolio, tell a collector’s story, or prove a curator’s exhibition history.

Given the higher demands of art world collections, we are long overdue for a fundamental shift in how NFT collections are managed. The current system has six severe limitations that we need to overcome to make collections universally useful: fragmentation, centralization, incompleteness, fragility, inconsistency, and the lack of a canonical identifier.

• Fragmentation occurs because each marketplace has its own collection system, leading to disconnected and proprietary formats. Collections are disconnected from related collections on other marketplaces, making it difficult to track which NFTs belong to a given collection or which collections a given NFT belongs to. This fragmentation limits the growth and richness of data relationships between collections.

• Centralization results from storing collections in marketplace-specific databases, risking data loss if platforms fail. This creates barriers between marketplaces, making it challenging to transfer or manage collections across multiple platforms due to the lack of a unified, portable data model.

• Incompleteness arises because current NFT collections often lack comprehensive metadata and contextual information, hindering the full representation of an artist’s work, a collector’s holdings, or a curator’s exhibition.

• Fragility is an almost natural property of web2 data. A database record can be deleted in an instant, and without a complete and persistent system of logs, it is next to impossible to verify—or even know—that the record ever existed.

• Inconsistency results from different platforms and databases recording and presenting information variably, leading to discrepancies that undermine the integrity of the collection and erode trust among stakeholders.

• The lack of a canonical identifier prevents referring to a collection as a whole. It cannot have an on-chain representation, nor is there a meaningful way of adding data or metadata that belongs to the whole collection. Additionally, without a standard identifier, there's no single source of truth for multiple digital copies of a collection.

Clearly, the current system is not a good foundation for building and maintaining long-lasting collections. Artists, collectors, and curators need a more reliable solution.

Artists frequently see their portfolio scattered across marketplaces. How should they present their work as a single collection? They have no choice but to manually assemble a “meta-collection” to catalog the various marketplace-specific collections. For example, the Brinkman NFT collection exists on 57 marketplaces, including SuperRare, NiftyGateway, Manifold, Rarible, and many more. The Brinkman NFT Catalog as a whole only exists as a table on Google Sheets, manually collected from all the marketplaces.But even when this tedious, manual work is done, and everything is gathered in one place, the collection is still not verifiable as a whole. While it is possible to verify each NFT in a collection individually, the exact contents of a collection remain the artist's unverified intention. Moreover, artists must frequently take action to keep their portfolio complete. Whenever a marketplace disappears, artists who have a collection listed there have to replicate it in another marketplace.

Collectors and curators have no choice but to trust the marketplace platforms. Unfortunately, the web3 idea of zero-trust transactions has yet to be applied to collections. The nature of web2 data makes verifying a collection’s authenticity, ownership, and completeness virtually impossible. The distribution of collections across marketplaces makes verifying any of these attributes even more difficult.

The only conclusion to be drawn from this flawed state of collections is this:

Collections must move onto the chain.

A canonical, on-chain definition of a collection would solve all of the existing deficiencies once and for all. Based on an on-chain collection, a universal collection standard could be established across marketplaces and chains.

Such a standardized, decentralized collection system would also enable essential features and functionalities of collections that the current system does not support. For example, at present, a tweet about an entire collection must be applied to all NFTs in a collection individually. With a decentralized collection, the tweet could be applied in one step.

To arrive at a world of verifiable collections, the NFT art world must find and agree on a universal collection format that covers all the needs of decentralized, verifiable collections.

For this purpose, Atomic Form proposes a system of attestations with a structured format to build a verifiable, immutable record of NFT collections.

The technical details of the proposal are beyond the scope of this article. In a nutshell, we propose a standard attestation and metadata structure for collections that works across chains and can be easily adopted by marketplaces via API providers. The core of an on-chain collection will be a collection ID that includes the chain, smart contract, and token ID to uniquely identify the collection. The structure is designed to cover different use cases, ranging from simple grouping of NFTs to more complex curation and management of collections.

The Brinkman Catalog example represents a simple use case: All NFTs exist on a single blockchain (Ethereum) but are spread across multiple marketplaces. This scenario presents no particular technical challenges, only organizational ones. All NFTs share the same structure, and the primary task here is to consolidate all the information from different marketplaces.

Transitioning single-chain NFT collections on-chain is, therefore, quite straightforward.

A much more complex use case is the Disobedient collection on OpenSea, with NFTs spread across five different blockchains. This scenario is quite involved on the technical level, as the collection must handle five different blockchain systems simultaneously, each with its own NFT standards. For example, unlike Ethereum, Solana doesn’t require a dedicated smart contract to mint NFTs; its built-in Token Program handles this.

Our on-chain collection proposal addresses these diverse requirements, offering the flexibility and scalability needed to manage multichain collections.

Several technical aspects have been considered when crafting the proposed structure, including:

• Handling unfinished collections: Collection owners must be able to add NFTs to a collection.

• Versioning and updates: All changes to a collection, such as transferring ownership, adding NFTs, or adding off-chain artifacts, should build up a traceable and verifiable history.

• Cross-chain compatibility: A collection should be able to contain NFTs from multiple chains.

• IPFS integration: Metadata must live in a content-addressable space to remain retrievable at any time.

These aspects have shaped the design, resulting in a universal and highly flexible solution.

What do NFT artists, collectors, and curators gain through this proposal?

First, collections become transparent and verifiable. By standardizing the collection structure and rooting it on a chain, everyone can inspect the contents and metadata of a collection.

Equally important is decentralization. Collections will no longer be at the mercy of a single company. Decentralized collections will be immune to marketplace shutdowns, securely living on countless copies of the chain and a decentralized storage system.

A standardized collection system also simplifies collection management. Instead of navigating the nuances of dozens of marketplace platforms, we will have a unified way to add NFTs to a collection, update lore, or transfer ownership —transparently and verifiably.

Finally, universal, chain-based collections provide opportunities for enhanced functionality that are not possible with the current system of isolated, marketplace-owned collections. Features like verifications, allow lists, airdrops, mass attestation via collection reference, or usage by AI agents become feasible, none of which work with off-chain, web2-based collections.

For example, an artist might decide to relocate a collection to another chain. Now, two copies of the collection exist. The same happens if a forger copies a collection. Being able to attest which instance is the correct one is crucial. Without a verifiable attestation, humans, algorithms, and AIs would have to rely on the artist’s website, tweets, or other unreliable web2 sources to tell the legitimate instance from the one that’s been forged or obsoleted.

Our proposal opens up a world of possibilities for artists, collectors, and curators. Collections will become much more than an entry in a proprietary database. On-chain collections fundamentally change the way we can work with collections or any verifiable grouping of NFTs.

By addressing the current limitations of NFT collections, we aim to enhance the creative possibilities of NFTs for artists, collectors, and curators alike. We envision a future where NFT curation is more accessible, verifiable, and interoperable across platforms and chains.

So let’s move collections where they belong: on-chain, with the same level of immutability and verifiability as the NFTs they contain.