What Is NFT Metadata? On-Chain vs Off-Chain Storage

NFT metadata is the structured data describing what an NFT actually is — its name, description, image reference, and any individual traits or attributes — and understanding where that metadata is stored matters because it determines how durable and tamper-resistant your NFT's actual content really is.

When people say they "own an NFT," what's recorded directly on the blockchain is typically just a token ID within a specific contract, plus a pointer (a URI) to where the associated metadata lives. The richness people actually care about — the artwork, the traits, the description — usually lives in that referenced metadata file, not in the on-chain record itself.

What a typical metadata file contains

A standard NFT metadata file (commonly formatted as JSON) typically includes:

• Name — the NFT's title.
• Description — text describing the item or artwork.
• Image — a reference (often another URI) pointing to the actual visual file.
• Attributes/traits — structured data like "background: blue," "hat: crown," used by marketplaces to display traits and by rarity tools to calculate rankings.
• Additional fields — depending on the project, this can include animation files, external links, or custom properties specific to that collection.

Fully on-chain vs. off-chain metadata storage

Fully on-chain metadata stores the entire metadata (and sometimes even the image itself, encoded directly into the data) within the smart contract or directly on the blockchain. This is the most durable and tamper-resistant option, since it inherits the blockchain's own persistence guarantees — as long as the chain exists, the metadata exists, with no dependency on any external server or pinning service.

Off-chain metadata stores the actual metadata file elsewhere — commonly on IPFS or Arweave, with only a reference URI recorded on-chain. This is dramatically cheaper in gas costs and is the far more common approach for larger collections, but it introduces a dependency on that off-chain storage remaining accessible over time.

Comparing storage approaches

	Fully on-chain	IPFS (off-chain, content-addressed)	Centralized server (off-chain, standard URL)
Gas cost to store	Very high, scales with data size	Low — only a reference hash stored on-chain	Low — only a URL stored on-chain
Durability	Extremely high — tied to blockchain itself	High if actively pinned; degrades if unpinned	Depends entirely on one company's server staying online
Tamper resistance	Very high	High — content hash changes if data is altered	Low — a company could alter or remove hosted files
Common use case	Small, simple collections; fully generative on-chain art	Most established NFT projects today	Some early or less carefully built projects

Why this distinction matters for value and durability

Two NFTs can look identical in a marketplace listing today but have very different long-term durability depending on where their metadata actually lives. An NFT with metadata on a centralized server controlled by a company that later shuts down could effectively lose its associated image and traits entirely — the token ID would still exist on-chain, but what it "points to" might return nothing. A project using IPFS with active pinning, or full on-chain storage, is considerably more resilient to that specific failure mode.

This is a genuine, practical due-diligence question worth asking about any NFT project: where does the metadata live, and who is responsible for keeping it accessible?

Metadata mutability: a further consideration

Some NFT contracts allow metadata to be updated after minting (useful for legitimate purposes like revealing traits after a blind mint period), while others lock metadata permanently at mint time. Mutable metadata isn't inherently bad, but it does mean the "same" NFT could theoretically point to different content over time if the contract allows the creator to change it — worth checking for any project where content permanence matters to you.

How to check a specific NFT's metadata setup yourself

Most block explorers allow you to look up the raw token URI recorded on-chain for a specific NFT, which reveals whether its metadata is referenced via IPFS, a similar decentralized storage network, or a standard centralized URL. This is a reasonable and fairly quick due-diligence step before a significant purchase — checking whether the metadata reference points somewhere durable, versus a single company's server that could go offline without warning.

Why this topic often gets overlooked

Most casual NFT buyers evaluate a piece based on how it looks and what marketplace displays about its traits, without checking where that information actually comes from or how durably it's stored. This is understandable — marketplaces generally render metadata seamlessly regardless of its underlying storage method, so the distinction is invisible during normal browsing. It only becomes visible, often unpleasantly, if underlying storage fails and a marketplace listing suddenly shows a broken image or missing metadata for an NFT someone already owns.

Bottom line

NFT metadata is the structured data — name, description, image reference, and traits — that gives an NFT its actual meaning, and where it's stored determines how durable that content really is over time. Fully on-chain storage offers the strongest guarantees but at high cost, while IPFS-based off-chain storage balances cost and durability reasonably well as long as content remains actively pinned. Centralized server-hosted metadata carries the most risk of the content simply disappearing. Learn more in our guide to IPFS for NFTs and our NFT glossary entry.