This answer covers a lot of ground, so it’s divided into three parts:
- How to use a CORS proxy to avoid “No Access-Control-Allow-Origin header” problems
- How to avoid the CORS preflight
- How to fix “Access-Control-Allow-Origin header must not be the wildcard” problems
How to use a CORS proxy to avoid “No Access-Control-Allow-Origin header” problems
If you don’t control the server your frontend code is sending a request to, and the problem with the response from that server is just the lack of the necessary
Access-Control-Allow-Origin header, you can still get things to work—by making the request through a CORS proxy.
You can easily run your own proxy with code from https://github.com/Rob–W/cors-anywhere/.
You can also easily deploy your own proxy to Heroku in just 2-3 minutes, with 5 commands:
git clone https://github.com/Rob--W/cors-anywhere.git cd cors-anywhere/ npm install heroku create git push heroku master
After running those commands, you’ll end up with your own CORS Anywhere server running at, e.g.,
Now, prefix your request URL with the URL for your proxy:
Adding the proxy URL as a prefix causes the request to get made through your proxy, which:
- Forwards the request to
- Receives the response from
- Adds the
Access-Control-Allow-Originheader to the response.
- Passes that response, with that added header, back to the requesting frontend code.
The browser then allows the frontend code to access the response, because that response with the
Access-Control-Allow-Origin response header is what the browser sees.
This works even if the request is one that triggers browsers to do a CORS preflight
OPTIONS request, because in that case, the proxy also sends the
Access-Control-Allow-Methods headers needed to make the preflight succeed.
How to avoid the CORS preflight
The code in the question triggers a CORS preflight—since it sends an
Even without that, the
Content-Type: application/json header will also trigger a preflight.
What “preflight” means: before the browser tries the
POST in the code in the question, it first sends an
OPTIONS request to the server, to determine if the server is opting-in to receiving a cross-origin
POST that has
Content-Type: application/json headers.
It works pretty well with a small curl script – I get my data.
To properly test with
curl, you must emulate the preflight
OPTIONS the browser sends:
curl -i -X OPTIONS -H "Origin: http://127.0.0.1:3000" \ -H 'Access-Control-Request-Method: POST' \ -H 'Access-Control-Request-Headers: Content-Type, Authorization' \ "https://the.sign_in.url"
https://the.sign_in.url replaced by whatever your actual
sign_in URL is.
The response the browser needs from that
OPTIONS request must have headers like this:
Access-Control-Allow-Origin: http://127.0.0.1:3000 Access-Control-Allow-Methods: POST Access-Control-Allow-Headers: Content-Type, Authorization
OPTIONS response doesn’t include those headers, the browser will stop right there and never attempt to send the
POST request. Also, the HTTP status code for the response must be a 2xx—typically 200 or 204. If it’s any other status code, the browser will stop right there.
The server in the question responds to the
OPTIONS request with a 501 status code, which apparently means it’s trying to indicate it doesn’t implement support for
OPTIONS requests. Other servers typically respond with a 405 “Method not allowed” status code in this case.
So you’ll never be able to make
OPTIONS request with a 405 or 501 or anything other than a 200 or 204 or if doesn’t respond with those necessary response headers.
The way to avoid triggering a preflight for the case in the question would be:
- if the server didn’t require an
Authorizationrequest header but instead, e.g., relied on authentication data embedded in the body of the
POSTrequest or as a query param
- if the server didn’t require the
POSTbody to have a
Content-Type: application/jsonmedia type but instead accepted the
application/x-www-form-urlencodedwith a parameter named
json(or whatever) whose value is the JSON data
How to fix “Access-Control-Allow-Origin header must not be the wildcard” problems
I am getting another error message:
The value of the ‘Access-Control-Allow-Origin’ header in the response
must not be the wildcard ‘*’ when the request’s credentials mode is
‘include’. Origin ‘
http://127.0.0.1:3000‘ is therefore not allowed
access. The credentials mode of requests initiated by the
XMLHttpRequest is controlled by the withCredentials attribute.
Access-Control-Allow-Origin header is
*. Instead the value in that case must exactly match your frontend code’s origin,
See Credentialed requests and wildcards in the MDN HTTP access control (CORS) article.
If you control the server you’re sending the request to, a common way to deal with this case is to configure the server to take the value of the
Origin request header, and echo/reflect that back into the value of the
Access-Control-Allow-Origin response header; e.g., with nginx:
add_header Access-Control-Allow-Origin $http_origin
But that’s just an example; other (web) server systems have similar ways to echo origin values.
I am using Chrome. I also tried using that Chrome CORS Plugin
That Chrome CORS plugin apparently just simplemindedly injects an
Access-Control-Allow-Origin: * header into the response the browser sees. If the plugin were smarter, what it would be doing is setting the value of that fake
So avoid using that plugin, even for testing. It’s just a distraction. To test what responses you get from the server with no browser filtering them, you’re better off using
curl -H as above.
fetch(…) request in the question:
headers.append('Access-Control-Allow-Origin', 'http://localhost:3000'); headers.append('Access-Control-Allow-Credentials', 'true');
Remove those lines. The
Access-Control-Allow-* headers are response headers. You never want to send them in requests. The only effect of that is to trigger a browser to do a preflight.