Managing NFT Metadata with IPFS

In this guide, you will rely on an IPFS workflow for ERC721 collections on Somnia.

This guide will walk you through how - You will prepare and upload artwork to IPFS via Pinata. - Generate clean, wallet/marketplace-friendly metadata JSON - Upload the metadata to IPFS via Pinata Cloud - Deploy a Solidity contract that stores the metadata URIs onchain (for each token) - Mint your NFTs.

All metadata resides on IPFS, while only the URIs (pointers) are stored onchain.

Please note that this guide provides a Smart Contract option for fully onchain metadata.

Concepts You Should Know

• IPFS (InterPlanetary File System) is a “Content Addressed Storage Network”. Files are addressed by their CID (content identifier). If any bit changes, the CID changes, and it is great for integrity and verifiability.

• Pinata: Pinata pins your files to IPFS and keeps them available. You’ll get CIDs (content identifiers) that never change for the same content.

• TokenURI: an ERC721 method that returns a URL (often an ipfs:// URI) pointing to a JSON file describing the NFT token (name, description, image, attributes, etc.).

• Per token URI vs baseURI:

  • Per token URI (this guide): store the full JSON URI onchain for each token with ERC721URIStorage. Easiest and most flexible.

  • baseURI pattern: compute tokenURI = baseURI + tokenId (no per token storage). Cheaper for large drops, but requires sequential naming.

Prerequisites

• MetaMask is configured for a Somnia RPC and has sufficient funds for gas.

• Node.js ≥ 18 (only if you’ll run the helper scripts below).

• Pinata account and a JWT (recommended): Get your JWT on Pinata

  • Pinata Dashboard → API Keys → New Key → choose Admin/Scoped as needed → copy JWT.

Prepare Images

If your art varies wildly in size or format, normalize once for a consistent user experience. Create an assets directory for your images. Run the node script below, targeting the assets directory to give uniformity to your images.

Install the dependencies:

npm i sharp fast-glob

Run the script:

// scripts/resize.js
import fg from "fast-glob";
import sharp from "sharp";
import { mkdirSync } from "fs";
import path from "path";

const INPUT = "assets/raw-images"; // put your original images here
const OUTPUT = "assets/images";

mkdirSync(OUTPUT, { recursive: true });

const files = (await fg(`${INPUT}/*.{png,jpg,jpeg,webp}`)).sort();
for (let i = 0; i < files.length; i++) {
  const out = path.join(OUTPUT, `${i}.png`);
  await sharp(files[i]).resize(1024, 1024, { fit: "cover" }).toFile(out);
}
console.log("Normalized images written to", OUTPUT);

Recommended Folder Structure

project/
  assets/
    images/
      0.png
      1.png
      2.png
      ...
    metadata/
      0.json
      1.json
      2.json
      ...
  scripts/
    resize.js              
    pinata.ts              (Pinata SDK init)
    upload-images.ts       (pin images folder → IMAGES_CID)
    make-metadata.ts       (generate JSON → uses IMAGES_CID)
    upload-metadata.ts     (pin metadata folder → METADATA_CID)

• images/ contain consistent dimensions and formats (e.g., 1024×1024 PNG).

• metadata/ contain one JSON per token ID, matching the filename (e.g., 7.json for tokenId 7).

Upload Images to IPFS via Pinata

Install Pinata SDK:

npm i pinata dotenv

Initialise Pinata using your JWT credentials:

// pinata.ts
import 'dotenv/config'
import { PinataSDK } from 'pinata'


export const pinata = new PinataSDK({
  pinataJwt: process.env.PINATA_JWT!,               // from Pinata “API Keys”
  pinataGateway: process.env.PINATA_GATEWAY!,       // e.g. myxyz.mypinata.cloud
})

Create a .env file in your project root:

PINATA_JWT=eyJhbGciOi...        # your Pinata JWT (keep secret)
PINATA_GATEWAY=myxyz.mypinata.cloud
IMAGES_CID=                     # leave blank until you upload images

Upload the images directory and get the root CID:

// scripts/upload-images.ts
import 'dotenv/config'
import { pinata } from './pinata'
import fs from 'node:fs/promises'
import path from 'node:path'
import { File } from 'node:buffer'

const DIR = 'assets/images'

async function main() {
  const names = (await fs.readdir(DIR))
    .filter(n => n.match(/\.(png|jpg|jpeg|webp)$/i))
    .sort((a, b) => Number(a.split('.')[0]) - Number(b.split('.')[0]))
    
  const files: File[] = []
  
  for (const name of names) {
    const bytes = await fs.readFile(path.join(DIR, name))
    files.push(new File([bytes], name)) // uploaded as a single folder
  }
  
  const res = await pinata.upload.public.fileArray(files)
  // res.cid is the directory root CID
  console.log('Images CID:', res.cid)
}

main().catch(console.error)

Run the following command to upload the images:

node dist/scripts/upload-images.js
# => Images CID: bafybe...

Copy the CID into .env as IMAGES_CID=....

Generate Metadata JSON

Each *.json should follow the de facto standards:

{
  "name": "NFTTest #0",
  "description": "A clean ERC-721 on Somnia with per-token metadata.",
  "image": "ipfs://IMAGES_CID/0.png",
  "external_url": "https://your-site.example",
  "attributes": [
    { "trait_type": "Edition", "value": 0 }
  ]
}

Install fast-glob library:

npm i fs-extra fast-glob

Run the script below to generate a file per image:

/ scripts/make-metadata.js
import fg from "fast-glob";
import { writeJSON, mkdirs } from "fs-extra";
import path from "path";


const IMAGES_CID = process.env.IMAGES_CID || "bafy..."; // paste from Step 2
const OUT_DIR = "assets/metadata";


sync function main() {
  await mkdirs(OUT_DIR)
  const imgs = (await fg('assets/images/*.{png,jpg,jpeg,webp}')).sort((a, b) =>
    Number(path.basename(a).split('.')[0]) - Number(path.basename(b).split('.')[0])
  )


  for (const p of imgs) {
    const id = Number(path.basename(p).split('.')[0])
    const json = {
      name: `NFTTest #${id}`,
      description: 'ERC-721 on Somnia with Pinata-hosted IPFS metadata.',
      image: `ipfs://${IMAGES_CID}/${id}.png`,
      attributes: [{ trait_type: 'Edition', value: id }]
    }
    await writeJSON(path.join(OUT_DIR, `${id}.json`), json, { spaces: 2 })
  }
  console.log('Metadata written to', OUT_DIR)
}
main().catch(console.error)

To create the metadata files, run the command:

node dist/scripts/make-metadata.js

Upload Metadata to IPFS

Upload the metadata folder to generate the METADATA_CID

// scripts/upload-metadata.ts
import { pinata } from './pinata'
import fs from 'node:fs/promises'
import path from 'node:path'
import { File } from 'node:buffer'

const DIR = 'assets/metadata'
async function main() {
  const names = (await fs.readdir(DIR))
    .filter(n => n.endsWith('.json'))
    .sort((a, b) => Number(a.split('.')[0]) - Number(b.split('.')[0]))
  const files: File[] = []
  for (const name of names) {
    const bytes = await fs.readFile(path.join(DIR, name))
    files.push(new File([bytes], name, { type: 'application/json' }))
  }
  const res = await pinata.upload.public.fileArray(files)
  console.log('Metadata CID:', res.cid)
}
main().catch(console.error)

Run the command:

node dist/scripts/upload-metadata.js
# => Metadata CID: bafybe...

Your tokenURI format is now: ipfs://<METADATA_CID>/<tokenId>.json Make sure the script uploads assets/metadata and records the printed Metadata CID; you’ll mint with: ipfs://<METADATA_CID>/<tokenId>.json

NFT Smart Contract

We’ll use ERC721URIStorage Smart Contract and mint with full URIs. Paste this into Remix as NFTTest.sol

NFTTest.sol

// SPDX-License-Identifier: MIT
// Compatible with OpenZeppelin Contracts ^5.4.0
pragma solidity ^0.8.27;


import {ERC721} from "@openzeppelin/contracts/token/ERC721/ERC721.sol";
import {ERC721URIStorage} from "@openzeppelin/contracts/token/ERC721/extensions/ERC721URIStorage.sol";
import {Ownable} from "@openzeppelin/contracts/access/Ownable.sol";

contract NFTTest is ERC721, ERC721URIStorage, Ownable {
    uint256 private _nextTokenId;

    constructor(address initialOwner)
        ERC721("NFTTest", "NFTT")
        Ownable(initialOwner)
    {}

    function _baseURI() internal pure override returns (string memory) {
        return "https://ipfs.io"; // only affects relative paths
    }

    function safeMint(address to, string memory uri)
        public
        onlyOwner
        returns (uint256)
    {
        uint256 tokenId = _nextTokenId++;
        _safeMint(to, tokenId);
        _setTokenURI(tokenId, uri); // store full URI (e.g., ipfs://CID/0.json)
        return tokenId;
    }

    function tokenURI(uint256 tokenId)
        public
        view
        override(ERC721, ERC721URIStorage)
        returns (string memory)
    {
        return super.tokenURI(tokenId);
    }


    function supportsInterface(bytes4 interfaceId)
        public
        view
        override(ERC721, ERC721URIStorage)
        returns (bool)
    {
        return super.supportsInterface(interfaceId);
    }
}

Contract Walkthrough

Imports

• ERC721 is the core NFT logic (ownership, transfers, approvals).

• ERC721URIStorage adds per-token storage for URIs via _setTokenURI.

• Ownable simple admin model; exposes onlyOwner for minting control.

ERC721URIStorage is the most straightforward way to store a token’s exact URI. For large drops, consider using a baseURI pattern to save storage gas; otherwise, this is perfect for flexible, explicit URIs.

State (_nextTokenId) is the sequential ID counter starting at 0. First mint → tokenId = 0, then 1, 2, … etc

Constructor

Initializes collection name/symbol and sets the owner to initialOwner (the only account allowed to mint).

_baseURI()

Returns https://ipfs.io. This is only used if you mint with relative paths (e.g., /ipfs/CID/7.json). If you mint with absolute ipfs://… URIs (recommended), this value is ignored.

safeMint(to, uri):

Owner-only mint. Creates a new token ID, mints safely (checking receiver contracts implement IERC721Receiver), and stores the exact uri string for that token via _setTokenURI.

Compile and Deploy using Remix. Follow this guide

Copy the deployed contract address.

How To Mint NFTs (Per Token URIs)

n Remix (Deployed Contracts):

• Call safeMint(to, uri) where:

  • to is the recipient address (can be yours).

  • uri is ipfs://<METADATA_CID>/<id>.json (e.g., ipfs://bafy.../0.json).

Then call tokenURI(0) to confirm the stored URI.

If you prefer scripts for multiple mints, you can write a script to Batch Mint using Hardhat, for example:

scripts/mint.ts (Hardhat)
import { ethers } from "hardhat";


const CONTRACT = "0xYourDeployedAddress";
const RECEIVER = "0xReceiver";
const METADATA_CID = "bafy...";


async function main() {
  const nft = await ethers.getContractAt("NFTTest", CONTRACT);
  for (let id = 0; id < 10; id++) {
    const uri = `ipfs://${METADATA_CID}/${id}.json`;
    const tx = await nft.safeMint(RECEIVER, uri);
    await tx.wait();
    console.log(`Minted #${id} → ${uri}`);
  }
}
main().catch(console.error);

Validation

To validate everything end to end, first do quick gateway smoke tests by opening https://ipfs.io/ipfs/<IMAGES_CID>/0.png and https://ipfs.io/ipfs/<METADATA_CID>/0.json in a browser to confirm the image and JSON are reachable. Next, call tokenURI(0) on your contract and verify it returns ipfs://<METADATA_CID>/0.json. Finally, check in a wallet or marketplace UI that the NFT renders correctly using the ipfs:// image URI embedded in the JSON.

MetadataOptional Fully Onchain Metadata

Smart Contract

If your art is small (SVG) or you only need text metadata, you can encode metadata JSON on-chain:

// SPDX-License-Identifier: MIT
pragma solidity ^0.8.27;

import {ERC721} from "@openzeppelin/contracts/token/ERC721/ERC721.sol";
import {Ownable} from "@openzeppelin/contracts/access/Ownable.sol";
import {Strings} from "@openzeppelin/contracts/utils/Strings.sol";
import {Base64} from "@openzeppelin/contracts/utils/Base64.sol";

contract OnchainMeta is ERC721, Ownable {
    using Strings for uint256;
    uint256 private _id;

    constructor(address owner_) ERC721("Onchain", "ONC") Ownable(owner_) {}

    function mint(address to) external onlyOwner returns (uint256) {
        uint256 tokenId = _id++;
        _safeMint(to, tokenId);
        return tokenId;
    }

    function tokenURI(uint256 tokenId) public view override returns (string memory) {
        _requireOwned(tokenId);
        // Example: trivial, static image via SVG data URI
        string memory image = string(
            abi.encodePacked(
                "data:image/svg+xml;base64,",
                Base64.encode(
                    bytes(
                        string(
                            abi.encodePacked(
                                "<svg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 512 512'>",
                                "<rect width='512' height='512' fill='black'/>",
                                "<text x='50%' y='50%' dominant-baseline='middle' text-anchor='middle' fill='white' font-size='48'>#",
                                tokenId.toString(),
                                "</text></svg>"
                            )
                        )
                    )
                )
            )
        );

        bytes memory json = abi.encodePacked(
            '{"name":"Onchain #', tokenId.toString(),
            '","description":"Fully on-chain metadata","image":"', image, '"}'
        );

        return string(abi.encodePacked(
            "data:application/json;base64,",
            Base64.encode(json)
        ));
    }
}

Trade-offs

• No external dependencies; URIs are immutable, always available.

• Higher gas (long strings), limited to compact media (SVG/text). For large images, prefer IPFS.