Finetune Images on Story

Completed Code

View the completed code for this tutorial.

In this tutorial, you will use the FLUX Finetuning API to take a bunch of images of Story’s mascot “Ippy” and finetune an AI model to create similar images along with a prompt. Then you will monetize and protect the output IP on Story.

This Tutorial is in TypeScript

Steps 1-3 of this tutorial are based on the FLUX Finetuning Beta Guide, which contains examples for calling their API in Python, however I have rewritten them in TypeScript.

The Explanation

Generative text-to-image models often do not fully capture a creator’s unique vision, and have insufficient knowledge about specific objects, brands or visual styles. With the FLUX Pro Finetuning API, creators can use existing images to finetune an AI to create similar images, along with a prompt.

When an image is created, we will register it as IP on Story in order to grow, monetize, and protect the IP.

0. Before you Start

There are a few steps you have to complete before you can start the tutorial.

You will need to install Node.js and npm. If you’ve coded before, you likely have these.
Add your Story Network Testnet wallet’s private key to .env file:

.env

WALLET_PRIVATE_KEY=

Go to Pinata and create a new API key. Add the JWT to your .env file:

.env

PINATA_JWT=

Go to BFL and create a new API key. Add the new key to your .env file:

BFL Credits

In order to generate an image, you’ll need BFL credits. If you just created an account, you will need to purchase credits here.

You can also see the pricing for each of the API endpoints here.

.env

BFL_API_KEY=

Add your preferred Story RPC URL to your .env file. You can just use the public default one we provide:

.env

RPC_PROVIDER_URL=https://aeneid.storyrpc.io

Install the dependencies:

Terminal

npm install @story-protocol/core-sdk axios pinata-web3 viem

1. Compile the Training Data

In order to create a finetune, we’ll need the input training data!

Create a folder in your project called images. In that folder, add a bunch of images that you want your finetune to train on. Supported formats: JPG, JPEG, PNG, and WebP. Also recommended to use more than 5 images.
Add Text Descriptions (Optional): In the same folder, create text files with descriptions for your images. Text files should share the same name as their corresponding images. Example: if your image is “sample.jpg”, create “sample.txt”
Compress your folder into a ZIP file. It should be named images.zip

2. Create a Finetune

In order to generate an image using a similar style as input images, we need to create a finetune. Think of a finetune as an AI that knows all of your input images and can then start producing new ones.

Let’s make a function that calls FLUX’s /v1/finetune API route. Create a flux folder, and inside that folder add a file named requestFinetuning.ts and add the following code:

Official Docs

In order to learn what each of the parameters in the payload are, see the official /v1/finetune API docs here.

flux/requestFinetuning.ts

import axios from "axios";
import fs from "fs";

interface FinetunePayload {
  finetune_comment: string;
  trigger_word: string;
  file_data: string;
  iterations: number;
  mode: string;
  learning_rate: number;
  captioning: boolean;
  priority: string;
  lora_rank: number;
  finetune_type: string;
}

export async function requestFinetuning(
  zipPath: string,
  finetuneComment: string,
  triggerWord = "TOK",
  mode = "general",
  iterations = 300,
  learningRate = 0.00001,
  captioning = true,
  priority = "quality",
  finetuneType = "full",
  loraRank = 32
) {
  if (!fs.existsSync(zipPath)) {
    throw new Error(`ZIP file not found at ${zipPath}`);
  }

  const modes = ["character", "product", "style", "general"];
  if (!modes.includes(mode)) {
    throw new Error(`Invalid mode. Must be one of: ${modes.join(", ")}`);
  }

  const fileData = fs.readFileSync(zipPath);
  const encodedZip = Buffer.from(fileData).toString("base64");

  const url = "https://api.us1.bfl.ai/v1/finetune";
  const headers = {
    "Content-Type": "application/json",
    "X-Key": process.env.BFL_API_KEY,
  };

  const payload: FinetunePayload = {
    finetune_comment: finetuneComment,
    trigger_word: triggerWord,
    file_data: encodedZip,
    iterations,
    mode,
    learning_rate: learningRate,
    captioning,
    priority,
    lora_rank: loraRank,
    finetune_type: finetuneType,
  };

  try {
    const response = await axios.post(url, payload, { headers });
    return response.data;
  } catch (error) {
    throw new Error(`Finetune request failed: ${error}`);
  }
}

Next, create a file named train.ts and call the requestFinetuning function we just made:

Warning: This is expensive!

Creating a new finetune is expensive, ranging from $2-$ 6 at the time of me writing this tutorial. Please review the “FLUX PRO FINETUNE: TRAINING” section on the pricing page.

train.ts

import { requestFinetuning } from "./flux/requestFinetuning";

async function main() {
  const response = await requestFinetuning("./images.zip", "ippy-finetune");
  console.log(response);
}

main();

This will log something that looks like:

{
  "finetune_id": "6fc5e628-6f56-48ec-93cb-c6a6b22bf5a"
}

This is your finetune_id, and will be used to create images in the following steps.

3. Wait for Finetune

Before we can generate images with our finetuned model, we have to wait for FLUX to finish training!

In our flux folder, create a file named finetune-progress.ts and add the following code:

Official Docs

In order to learn what each of the parameters in the payload are, see the official /v1/get_result API docs here.

flux/finetuneProgress.ts

import axios from "axios";

export async function finetuneProgress(finetuneId: string) {
  const url = "https://api.us1.bfl.ai/v1/get_result";
  const headers = {
    "Content-Type": "application/json",
    "X-Key": process.env.BFL_API_KEY,
  };
  try {
    const response = await axios.get(url, {
      headers,
      params: { id: finetuneId },
    });
    return response.data;
  } catch (error) {
    throw new Error(`Finetune progress failed: ${error}`);
  }
}

Next, create a file named finetune-progress.ts and call the finetuneProgress function we just made:

finetune-progress.ts

import { finetuneProgress } from "./flux/finetuneProgress";

// input your finetune_id here
const FINETUNE_ID = "";

async function main() {
  const progress = await finetuneProgress(FINETUNE_ID);
  console.log(progress);
}

main();

This will log something that looks like:

{
  "id": "023a1507-369e-46e0-bd6d-1f3446d7d5f2",
  "status": "Pending",
  "result": null,
  "progress": null
}

As you can see, the status is still pending. We must wait until the training is ‘Ready’ before we can move on to the next step.

4. Run Inference

Warning: This costs money.

Although very cheap, running an inference does cost money, ranging from $0.06-0.07 at the time of me writing this tutorial. Please review the “FLUX PRO FINETUNE: INFERENCE” section on the pricing page.

Now that we have trained a finetune, we will use the model to create images. “Running an inference” simply means using our new model (identified by its finetune_id), which is trained on our images, to create new images.

There are several different inference endpoints we can use, each with their own pricing (found at the bottom of the page). For this tutorial, I’ll be using the /v1/flux-pro-1.1-ultra-finetuned endpoint, which is documented here.

In our flux folder, create a finetuneInference.ts file and add the following code:

Official Docs

In order to learn what each of the parameters in the payload are, see the official /v1/flux-pro-1.1-ultra-finetuned API docs here.

flux/finetineInference.ts

import axios from "axios";

export async function finetuneInference(
  finetuneId: string,
  prompt: string,
  finetuneStrength = 1.2,
  endpoint = "flux-pro-1.1-ultra-finetuned",
  additionalParams: Record<string, any> = {}
) {
  const url = `https://api.us1.bfl.ai/v1/${endpoint}`;
  const headers = {
    "Content-Type": "application/json",
    "X-Key": process.env.BFL_API_KEY,
  };

  const payload = {
    finetune_id: finetuneId,
    finetune_strength: finetuneStrength,
    prompt,
    ...additionalParams,
  };

  try {
    const response = await axios.post(url, payload, { headers });
    return response.data;
  } catch (error) {
    throw new Error(`Finetune inference failed: ${error}`);
  }
}

Next, create a file named inference.ts and call the finetuneInference function we just made. The first parameter should be the finetune_id we got from running the script above, and the second parameter is a prompt to generate a new image.

inference.ts

import { finetuneInference } from "./flux/finetuneInference";

// input your finetune_id here
const FINETUNE_ID = "";
// add your prompt here
const PROMPT = "A picture of Ippy being really happy.";

async function main() {
  const inference = await finetuneInference(FINETUNE_ID, PROMPT);
  console.log(inference);
}

main();

This will log something that looks like:

{
  "id": "023a1507-369e-46e0-bd6d-1f3446d7d5f2",
  "status": "Pending",
  "result": null,
  "progress": null
}

As you can see, the status is still pending. We must wait until the generation is ready to view our image. To do this, we will need a function to fetch our new inference to see if its ready and view the details about it.

In our flux folder, create a file named getInference.ts and add the following code:

Official Docs

In order to learn what each of the parameters in the payload are, see the official /v1/get_result API docs here.

flux/getInference.ts

import axios from "axios";

export async function getInference(id: string) {
  const url = "https://api.us1.bfl.ai/v1/get_result";
  const headers = {
    "Content-Type": "application/json",
    "X-Key": process.env.BFL_API_KEY,
  };

  try {
    const response = await axios.get(url, { headers, params: { id } });
    return response.data;
  } catch (error) {
    throw new Error(`Inference retrieval failed: ${error}`);
  }
}

Back in our inference.ts file, lets add a loop that continuously fetches the inference until it’s ready. When it’s ready, we will view the new image.

inference.ts

import { finetuneInference } from "./flux/finetuneInference";
import { getInference } from "./flux/getInference";

// input your finetune_id here
const FINETUNE_ID = "";
// add your prompt here
const PROMPT = "A picture of Ippy being really happy.";

async function main() {
  const inference = await finetuneInference(FINETUNE_ID, PROMPT);

  let inferenceData = await getInference(inference.id);
  while (inferenceData.status != "Ready") {
    console.log("Waiting for inference to complete...");
    // wait 5 seconds
    await new Promise((resolve) => setTimeout(resolve, 5000));
    // fetch the inference again
    inferenceData = await getInference(inference.id);
  }
  // now the inference data is ready
  console.log(inferenceData);
}

main();

Once the loop completed, the final log will look like:

{
  "id": "023a1507-369e-46e0-bd6d-1f3446d7d5f2",
  "status": "Ready",
  "result": {
    "sample": "https://delivery-us1.bfl.ai/results/746585f8d1b341f3a8735ababa563ac1/sample.jpeg?se=2025-01-16T19%3A50%3A11Z&sp=r&sv=2024-11-04&sr=b&rsct=image/jpeg&sig=pPtWnntLqc49hfNnGPgTf4BzS6MZcBgHayrYkKe%2BZIc%3D",
    "prompt": "A picture of Ippy being really happy."
  },
  "progress": null
}

You can paste the sample into your browser and see the final result! Make sure to save this image as it will disappear eventually.

5. Set up your Story Config

Next we will register this image on Story as an IP Asset in order to monetize and license the IP. Create a story folder and add a utils.ts file. In there, add the following code to set up your Story Config:

Associated docs: TypeScript SDK Setup

story/utils.ts

import { StoryClient, StoryConfig } from "@story-protocol/core-sdk";
import { http } from "viem";
import { privateKeyToAccount, Address, Account } from "viem/accounts";

const privateKey: Address = `0x${process.env.WALLET_PRIVATE_KEY}`;
export const account: Account = privateKeyToAccount(privateKey);

const config: StoryConfig = {
  account: account,
  transport: http(process.env.RPC_PROVIDER_URL),
  chainId: "aeneid",
};
export const client = StoryClient.newClient(config);

6. Upload Inference to IPFS

Now that we have made a new inference, we’ll have to store the image sample file ourselves on IPFS because the sample is only temporary.

In a new pinata folder, create a uploadToIpfs.ts file and create a function to upload our image and get details about it:

pinata/uploadToIpfs.ts

import { PinataSDK } from "pinata-web3";

const pinata = new PinataSDK({
  pinataJwt: process.env.PINATA_JWT,
});

export async function uploadImageAndGetDetails(
  imageUrl: string
): Promise<{ ipfsCid: string; contentType: string; contentHash: string }> {
  try {
    const response = await axios.get(imageUrl, {
      responseType: "arraybuffer",
      validateStatus: (status) => status === 200,
    });

    const contentType = response.headers["content-type"];
    if (!contentType?.startsWith("image/")) {
      throw new Error("URL does not point to an image");
    }

    const extension = contentType.split("/")[1];
    const filename = `${Date.now()}-${Math.random()
      .toString(36)
      .slice(2)}.${extension}`;

    const buffer = Buffer.from(response.data);
    const contentHash =
      "0x" + createHash("sha256").update(buffer).digest("hex");
    const file = new File([buffer], filename, { type: contentType });

    const { IpfsHash } = await pinata.upload.file(file);
    return { ipfsCid: IpfsHash, contentType, contentHash };
  } catch (error) {
    if (axios.isAxiosError(error)) {
      throw new Error(`Failed to fetch image: ${error.message}`);
    }
    throw error;
  }
}

We will now use this function in the following step.

7. Set up your IP Metadata

In your story folder, create a registerIp.ts file.

View the IPA Metadata Standard and construct the metadata for your IP as shown below:

story/registerIp.ts

import { client, account } from "./utils";
import { uploadImageAndGetDetails } from "../pinata/uploadToIpfs";

export async function registerIp(inference) {
  const { ipfsCid, contentType, contentHash } = await uploadImageAndGetDetails(
    inference.result.sample
  );

  // Docs: https://docs.story.foundation/concepts/ip-asset/ipa-metadata-standard
  const ipMetadata = {
    title: "Happy Ippy",
    description:
      "An image of Ippy being really happy, generated by FLUX's 1.1 [pro] ultra Finetune",
    image: `https://ipfs.io/ipfs/${ipfsCid}`,
    imageHash: contentHash,
    mediaUrl: `https://ipfs.io/ipfs/${ipfsCid}`,
    mediaHash: contentHash,
    mediaType: contentType,
    creators: [
      {
        name: "Jacob Tucker",
        contributionPercent: 100,
        address: account.address,
      },
    ],
  };
}

8. Set up your NFT Metadata

In the registerIp.ts file, configure your NFT Metadata, which follows the OpenSea ERC-721 Standard.

story/registerIp.ts

import { client, account } from "./utils";
import { uploadImageAndGetDetails } from "../pinata/uploadToIpfs";

export async function registerIp(inference) {
  // previous code here...

  // Docs: https://docs.opensea.io/docs/metadata-standards
  const nftMetadata = {
    name: "Ippy Ownership NFT",
    description:
      "This NFT represents ownership of the Happy Ippy image generated by FLUX's 1.1 [pro] ultra Finetune",
    image: `https://ipfs.io/ipfs/${ipfsCid}`,
    attributes: [
      {
        key: "Model",
        value: "FLUX 1.1 [pro] ultra Finetune",
      },
      {
        key: "Prompt",
        value: "A picture of Ippy being really happy.",
      },
    ],
  };
}

9. Upload your IP and NFT Metadata to IPFS

In the pinata folder, create a function to upload your IP & NFT Metadata objects to IPFS:

pinata/uploadToIpfs.ts

// previous code here ...

export async function uploadJSONToIPFS(jsonMetadata: any): Promise<string> {
  const { IpfsHash } = await pinata.upload.json(jsonMetadata);
  return IpfsHash;
}

You can then use that function to upload your metadata, as shown below:

story/registerIp.ts

import { client, account } from "./utils";
import {
  uploadImageAndGetDetails,
  uploadJSONToIPFS,
} from "../pinata/uploadToIpfs";
import { createHash } from "crypto";

export async function registerIp(inference) {
  // previous code here...

  const ipIpfsHash = await uploadJSONToIPFS(ipMetadata);
  const ipHash = createHash("sha256")
    .update(JSON.stringify(ipMetadata))
    .digest("hex");
  const nftIpfsHash = await uploadJSONToIPFS(nftMetadata);
  const nftHash = createHash("sha256")
    .update(JSON.stringify(nftMetadata))
    .digest("hex");
}

10. Register the NFT as an IP Asset

Next we will mint an NFT, register it as an IP Asset, set License Terms on the IP, and then set both NFT & IP metadata.

Luckily, we can use the mintAndRegisterIp function to mint an NFT and register it as an IP Asset in the same transaction.

This function needs an SPG NFT Contract to mint from. For simplicity, you can use a public collection we have created for you on Aeneid testnet: 0xc32A8a0FF3beDDDa58393d022aF433e78739FAbc.

Creating your own custom ERC-721 collection

Using the public collection we provide for you is fine, but when you do this for real, you should make your own NFT Collection for your IPs. You can do this in 2 ways:

Deploy a contract that implements the ISPGNFT interface, or use the SDK’s createNFTCollection function (shown below) to do it for you. This will give you your own SPG NFT Collection that only you can mint from.

createSpgNftCollection.ts

import { zeroAddress } from "viem";
import { client } from "./utils";

async function createSpgNftCollection() {
  const newCollection = await client.nftClient.createNFTCollection({
    name: "Test NFTs",
    symbol: "TEST",
    isPublicMinting: false,
    mintOpen: true,
    mintFeeRecipient: zeroAddress,
    contractURI: "",
  });

  console.log(
    `New SPG NFT collection created at transaction hash ${newCollection.txHash}`
  );
  console.log(`NFT contract address: ${newCollection.spgNftContract}`);
}

createSpgNftCollection();

Create a custom ERC-721 NFT collection on your own and use the register function - providing an nftContract and tokenId - instead of using the mintAndRegisterIp function. See a working code example here. This is helpful if you already have a custom NFT contract that has your own custom logic, or if your IPs themselves are NFTs.

Associated Docs: ipAsset.mintAndRegisterIp

story/registerIp.ts

import { client, account } from "./utils";
import {
  uploadImageAndGetDetails,
  uploadJSONToIPFS,
} from "../pinata/uploadToIpfs";
import { createHash } from "crypto";
import { Address } from "viem";

export async function registerIp(inference) {
  // previous code here ...

  const response = await client.ipAsset.mintAndRegisterIp({
    spgNftContract: "0xc32A8a0FF3beDDDa58393d022aF433e78739FAbc",
    ipMetadata: {
      ipMetadataURI: `https://ipfs.io/ipfs/${ipIpfsHash}`,
      ipMetadataHash: `0x${ipHash}`,
      nftMetadataURI: `https://ipfs.io/ipfs/${nftIpfsHash}`,
      nftMetadataHash: `0x${nftHash}`,
    },
  });

  console.log(
    `Root IPA created at transaction hash ${response.txHash}, IPA ID: ${response.ipId}`
  );
  console.log(
    `View on the explorer: https://aeneid.explorer.story.foundation/ipa/${response.ipId}`
  );
}

11. Register our Inference

Now that we have completed our registerIp function, let’s add it to our inference.ts file:

inference.ts

import { finetuneInference } from "./flux/finetuneInference";
import { getInference } from "./flux/getInference";
import { registerIp } from "./story/registerIp";

const FINETUNE_ID = "";
const PROMPT = "A picture of Ippy being really happy.";

async function main() {
  const inference = await finetuneInference(FINETUNE_ID, PROMPT);

  let inferenceData = await getInference(inference.id);
  while (inferenceData.status != "Ready") {
    console.log("Waiting for inference to complete...");
    await new Promise((resolve) => setTimeout(resolve, 5000));
    inferenceData = await getInference(inference.id);
  }
  // now the inference data is ready
  console.log(inferenceData);

  // add the function here
  await registerIp(inferenceData);
}

main();

Get Started

Integration Guides

Tutorials

Infrastructure Partners

Completed Code

The Explanation

0. Before you Start

1. Compile the Training Data

2. Create a Finetune

3. Wait for Finetune

4. Run Inference

5. Set up your Story Config

6. Upload Inference to IPFS

7. Set up your IP Metadata

8. Set up your NFT Metadata

9. Upload your IP and NFT Metadata to IPFS

10. Register the NFT as an IP Asset

11. Register our Inference

12. Done!

Completed Code

Learn More

Get Started

Integration Guides

Tutorials

Infrastructure Partners

Completed Code

​The Explanation

​0. Before you Start

​1. Compile the Training Data

​2. Create a Finetune

​3. Wait for Finetune

​4. Run Inference

​5. Set up your Story Config

​6. Upload Inference to IPFS

​7. Set up your IP Metadata

​8. Set up your NFT Metadata

​9. Upload your IP and NFT Metadata to IPFS

​10. Register the NFT as an IP Asset

​11. Register our Inference

​12. Done!

Completed Code

Learn More

The Explanation

0. Before you Start

1. Compile the Training Data

2. Create a Finetune

3. Wait for Finetune

4. Run Inference

5. Set up your Story Config

6. Upload Inference to IPFS

7. Set up your IP Metadata

8. Set up your NFT Metadata

9. Upload your IP and NFT Metadata to IPFS

10. Register the NFT as an IP Asset

11. Register our Inference

12. Done!