How to upload files to s3 in React using pre signed urls


Borislav Hadzhiev

Thu Apr 22 20218 min read

Updated on Thu Apr 22 2021

Upload files to an S3 bucket in a React.js application using AWS CDK to provision the infrastructure with a single command

Uploading Files with Pre Signed URLs in React #

In this article we are going to upload files to an S3 bucket in a React.js application using presigned urls.

We are going to deploy a CDK stack that will provision the following resources:

  • S3 bucket that will store our uploads
  • Api gateway with Lambda integration
  • Lambda function that creates and returns presigned urls to our React.js frontend

The flow is going to be:

  1. The frontend makes a request to the API gateway endpoint
  2. The Api endpoint has Lambda integration. The lambda function makes a request to S3 to get a presigned url, which is then returned to the frontend
  3. The frontend uses the presigned url to upload a file to s3

Prerequisites #

  • Have the AWS CLI installed and configured.

Project setup #

  1. Clone the github repository

  2. Install the dependencies

npm run setup
  1. Create the CDK stack
npm run cdk-create-stack
  1. At this point our stack, named presigned-url-dev is created and you can view its resources in the Cloudformation console.

  2. Start the react application. Note: it's very important that you run the react application on http://localhost:3000, because that's the url we've set up CORS for.

npm run dev

You can view the region where the stack was created, as well as the bucket name by opening the src/cdk-exports-dev.json file. This is where we store the outputs from the CDK stack creation.

Let's now test the application, initially we see a File input and an Upload to s3 button:

initial app state

Click on the File Input, select an image of up to 1 MB size and click on the Upload to s3 button to upload the image:

with upload

The image will be rendered below the Upload to s3 button. At this point we know our application works, so let's go over the moving parts.

Backend code #

We'll start with the backend, open up backend/infra/app.ts. This is the entry point of our application. Here we initialize our PresignedUrlStack.

new PresignedUrlStack(app, `${STACK_PREFIX}-${DEPLOY_ENVIRONMENT}`, {
  env: {
    region: DEPLOY_REGION,
  tags: {
    env: 'dev',

Let's look at the contents of the stack at backend/infra/presigned-url-stack.ts.

The first resource we create is the s3 bucket which will store our uploads:

const s3Bucket = new s3.Bucket(this, id, {
  cors: [
      allowedMethods: [
      allowedOrigins: [FRONTEND_BASE_URL],
      allowedHeaders: ['*'],

We've set up cors for the bucket, because we're going to access it from our frontend at http://localhost:3000.

Setting the CORS configuration is important when you access resources from a different domain. Since the s3 bucket is going to be hosted on the domain and our frontend is currently hosted at http://localhost:3000, these are two different domains and we must set up CORS.

The next resource we create is the API:

const httpApi = new apiGateway.HttpApi(this, 'api', {
  description: `___${DEPLOY_ENVIRONMENT}___ Api for ${STACK_PREFIX}`,
  corsPreflight: {
    allowHeaders: ['Content-Type', 'X-Amz-Date', 'Authorization', 'X-Api-Key'],
    allowMethods: [
    allowCredentials: true,
    allowOrigins: [FRONTEND_BASE_URL],

Again nothing special about the API configuration. The majority of the code is again the CORS configuration. We have to add our frontend url to the allowed origins so it can access our API, which is on a different domain.

The next resource we define is the Lambda function, which is responsible for talking to S3 and getting a presigned URL:

const getPresignedUrlFunction = new NodejsFunction(this, 'get-presigned-url', {
  runtime: lambda.Runtime.NODEJS_14_X,
  memorySize: 1024,
  timeout: cdk.Duration.seconds(5),
  handler: 'main',
  entry: path.join(__dirname, '/../src/get-presigned-url-s3/index.ts'),
  environment: {BUCKET_NAME: s3Bucket.bucketName},


We pass in the bucket name as an environment variable, because we'll make us of it in the lambda code.

A very important part is granting permissions to the Lambda to PUT objects in the s3 bucket and to modify the Access control list of objects in the bucket.

These are the permissions we need for the creation of presigned urls. If the Lambda doesn't have those permissions, you would get a 403 Unauthorized error when you try to upload a file from your frontend, because the presigned url would be invalid.

The last thing we do for our infrastructure is to add the Lambda function as an integration for the API GET route at /get-presigned-url-s3:

  path: '/get-presigned-url-s3',
  methods: [apiGateway.HttpMethod.GET],
  integration: new apiGatewayIntegrations.LambdaProxyIntegration({
    handler: getPresignedUrlFunction,

Now that we're done with the infrastructure let's move to the code, open up backend/src/get-presigned-url-s3/index.ts.

Our main function is the entry of our lambda. We first check for the fileType query parameter and if it wasn't provided we throw an error:

if (!event.queryStringParameters?.fileType)
  throw new Error(
    'Querystring parameter fileType must be provided when creating a presigned URL, i.e. ?fileType=image/png',

We are going to use the fileType query parameter to force the frontend into specifying the exact type of image the user intends to upload. That way we can constrain the presigned url to only image/png or image/jpeg, etc.

The next thing we do is generate a random name for the image:

const filePath = generateId();

function generateId() {
  let result = '';
  const characters =
  const length = 10;

  const charactersLength = characters.length;
  for (let i = 0; i < length; i += 1) {
    result += characters.charAt(Math.floor(Math.random() * charactersLength));

  const date = new Date().toISOString().split('T')[0].replace(/-/g, '');
  return `${date}_${result}`;

The generateId function generates a random name for the file. We don't want users to be able to overwrite the images uploaded by others, so we just randomize the name. You could use a uuid function or change the implementation to something more robust.

The next thing we do in our main function is to call the createPresignedPost function passing in the fileType and the generated name:

const presignedPost = await createPresignedPost({fileType, filePath});

export function createPresignedPost({
}: GetPresignedPostUrlParams): Promise<S3.PresignedPost> {
  const params = {
    Bucket: process.env.BUCKET_NAME,
    Fields: {key: filePath, acl: 'public-read'},
    Conditions: [
      // content length restrictions: 0-1MB]
      ['content-length-range', 0, 1000000],
      // specify content-type to be more generic- images only
      // ['starts-with', '$Content-Type', 'image/'],
      ['eq', '$Content-Type', fileType],
    // number of seconds for which the presigned policy should be valid
    Expires: 15,

  const s3 = new S3();
  return (s3.createPresignedPost(
  ) as unknown) as Promise<S3.PresignedPost>;

The createPresignedPost function talks to s3 and grabs the presigned url that we have to then return to our frontend.

Notice that we set the acl to public-read, which means that all users on the internet are able to view this file.

We then set certain conditions for the presigned url. The content length range restriction is set to 1 Megabyte, which means that we only allow users to upload files of up to 1 Megabyte.

Then we set a condition for the Content-Type of the file, it must be equal to the provided from the frontend fileType query parameter.

Finally we set the presigned url to expire in 15 seconds. The presigned url can be used at most once. Based on our configuration if it's not used within 15 seconds it automatically expires and becomes invalid.

We have covered all of our backend code, at this point we have all our infrastructure in place and a Lambda function, which is able to talk to s3 and generate presigned urls.

Frontend code #

Let's now look at the Frontend code, open src/pages/, which is the / route of our application.

Since we are going to be uploading images to s3, we know we'll need an input type='file'. The first lines of code in our Home component make use of the useFileChange hook:

const {
} = useFileChange();

The handleFileChange function is the onChange handler for our input type file. Let's take a look at what it does. It is defined in the src/pages/use-file-change.ts file:

const handleFileChange = (event: React.ChangeEvent<HTMLInputElement>) => {
  const fileObj = &&[0];
  if (!fileObj) {

  console.log('fileObj is', fileObj);

  const [type] = fileObj.type.split('/');
  if (!type || type !== 'image') {
      type: 'FILE_CHANGE_FAILURE',
      fileError: 'You can only upload image files.',

  if (fileObj.size > MAX_FILE_SIZE_BYTES) {
      type: 'FILE_CHANGE_FAILURE',
      fileError: `File is too large, file size is ${bytesToMb(
      ).toFixed(2)} MB, maximum allowed size - 1 MB.`,

  // eslint-disable-next-line no-param-reassign = '';

    fileSize: fileObj.size,
    fileType: fileObj.type,
    fileContents: fileObj,

It first grabs the first file from the event object, and then it checks whether the file is of type image. If the file is not an image we display an error to the user that they should only upload image files.

Next, we check if the file the user selected is larger than 1 Megabyte, if it is display an error message that they should upload files of up to 1 Megabyte.

We know that if the user were to upload a file larger than 1 Megabyte they would get an error because we've specified a condition for our presigned url, but we're placing this check just to shorten the feedback loop.

Lastly if the file is an image of size less than 1 megabyte, we dispatch a success event.

Next let's look at the submit handler, open src/pages/

const handleSubmit = async (e: FormEvent<HTMLFormElement>) => {

  try {
    if (fileType && fileContents) {
      const filePath = await uploadToS3({fileType, fileContents});
      console.log('filePath is', filePath);
      fileDispatch({type: 'RESET_FILE_STATE'});
  } catch (err) {
    console.log('error is', err);

In our submit handler we first call the uploadToS3 function, located in src/pages/upload-file.ts:

export async function uploadToS3({
}: {
  fileType: string;
  fileContents: File;
}) {
  const presignedPostUrl = await getPresignedPostUrl(fileType);

  const formData = new FormData();
  formData.append('Content-Type', fileType);
  Object.entries(presignedPostUrl.fields).forEach(([k, v]) => {
    formData.append(k, v);
  formData.append('file', fileContents); // The file has be the last element

  const response = await, formData, {
    headers: {'Content-Type': 'multipart/form-data'},

  return presignedPostUrl.filePath;

type PresignedPostUrlResponse = {
  url: string;
  fields: {
    key: string;
    acl: string;
    bucket: string;
  filePath: string;

const GET_PRESIGNED_URL_API_PATH = 'get-presigned-url-s3';

async function getPresignedPostUrl(fileType: string) {
  const {data: presignedPostUrl} = await axios.get<PresignedPostUrlResponse>(

  return presignedPostUrl;

The uploadToS3 function first gets the presigned url by making a get request to our backend API.

Then it uses FormData to construct a set of key value pairs representing the fields and values from the presigned url. Note that the file must be the last element of the FormData object.

Then the function makes a post request to the presigned url passing in the FormData and setting the Content-Type as multipart/form-data.

Going back to src/page/ our submit handler sets the s3 image url in the state, so we can render it:


Every time the file input changes the file will be set in the state, and every time the submit button is clicked we get a presigned url from our API and then use it to upload an image to s3.

Upload image success

You can check the contents of your s3 bucket by opening the src/cdk-exports-dev.json file to get the name of the bucket and then selecting the bucket in the AWS S3 console.

Upload image success

Cleanup #

npm run cdk-destroy

Note that in order to delete the s3 bucket you have to first empty its contents and then delete it.

Conclusion #

Let's summarize some of the things we have to be mindful of when using presigned urls:

  • The s3 bucket must have cors enabled, for us to be able to upload files from a web application, hosted on a different domain.
  • The lambda function that talks to s3 to get the presigned url must have permissions for s3:PutObject and s3:PutObjectAcl on the bucket.
  • To make the uploaded files publicly readable, we have to set the acl to public-read:
const params = {
  Bucket: bucketName,
  Fields: {
    key: filePath,
    acl: 'public-read',
  Conditions: [
    // content length restrictions: 0-1MB]
    ['content-length-range', 0, 1000000],
    // specify content-type to be more generic - images only
    // ['starts-with', '$Content-Type', 'image/'],
    ['eq', '$Content-Type', fileType],
    ['starts-with', '$key', identityId],
  // number of seconds for which the presigned policy should be valid
  Expires: 15,
  • The conditions that we specify must be met by the frontend, i.e. the file size, otherwise an error is thrown.

  • The default expiration for presigned urls is 15 minutes, but you can set a shorter one by passing in the Expires key.

Further Reading #

Join my newsletter

I'll send you 1 email a week with links to all of the articles I've written that week

Buy Me A Coffee