diff --git a/README.md b/README.md
index 9931f63d9..5cf465d91 100644
--- a/README.md
+++ b/README.md
@@ -10,6 +10,7 @@ GlusterD, while also becoming more modular and easing extensibility.
 ## Documentation
 
 * [Quick Start User Guide](doc/quick-start-user-guide.md)
+* [REST API Authentication](doc/rest-authentication.md)
 * [Development Guide](doc/development-guide.md)
 * [Coding Guidelines](doc/coding.md)
 * [REST API Reference](doc/endpoints.md)
diff --git a/doc/quick-start-user-guide.md b/doc/quick-start-user-guide.md
index 4163e9760..edc9d52a0 100644
--- a/doc/quick-start-user-guide.md
+++ b/doc/quick-start-user-guide.md
@@ -94,6 +94,8 @@ Now you have two nodes running glusterd2.
 
 > NOTE: Ensure that firewalld is configured (or stopped) to let traffic on ports ` before adding a peer.
 
+REST API authentication is enabled by default, read more about it [here](rest-authentication.md).
+
 ## Add peer
 
 Glusterd2 natively provides only ReST API for clients to perform management operations. A CLI is provided which interacts with glusterd2 using the [ReST APIs](https://github.com/gluster/glusterd2/wiki/ReST-API).
diff --git a/doc/rest-authentication.md b/doc/rest-authentication.md
new file mode 100644
index 000000000..be4e81ef3
--- /dev/null
+++ b/doc/rest-authentication.md
@@ -0,0 +1,106 @@
+# Glusterd2 REST API authentication
+
+Glusterd2 REST API authentication is based on [JWT](https://jwt.io). If REST
+authentication is enabled then each client request should include
+`Authorization` header as example below.
+
+    Authorization: bearer eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnbHVzdGVyY2xpIiwiaWF0IjoxNTMxNzI5NzQyLCJleHAiOjE1MzE3Mjk3NTJ9._WIwO7PrHUHIT62SdzfkyNjqD1GEgX2cYqN8ACZCtaw
+
+**Note**: REST Authentication can be disabled by adding `restauth=false`
+in Glusterd2 config file(Default path is
+`/var/lib/glusterd2/glusterd2.toml` in case of rpm installation)
+
+## Default user
+
+When `glusterd2` starts for the first time, it creates the
+`$GLUSTERD2_STATEDIR/auth` file which will contain the secret. If
+"gluster" user group is available in the system then this `auth` file
+can be read by any user in that machine who are part of "gluster"
+group.
+
+**Note**: If "gluster" user group is not available during first start
+of `glusterd2` it limits the read permission to `root:root`
+
+## glustercli
+
+With default installation, `glustercli` will know the location of
+`auth` file. `glustercli` will generate JWT token using the secret
+available in `auth` file and attach it with every REST API calls.
+
+Default installation will not require any change to use `glustercli`.
+
+If `auth` file is in different path(When `glusterd2` is running with
+custom `workdir`), then run `glustercli` by specifying
+`--secret-file`. For example,
+
+    glustercli --secret-file=/root/setup1/glusterd2/auth peer status
+
+**Note**: bash alias can be added like `alias glustercli='glustercli
+--secret-file=/root/setup1/glusterd2/auth`
+
+Secret is taken by `glustercli` in following order of precedence
+(highest to lowest):
+
+    --secret
+    --secret-file
+    GLUSTERD2_AUTH_SECRET (environment variable)
+    --secret-file (default path)
+
+## Curl example
+
+Download the utility script `glustercli-auth-header.py` from
+[here](https://github.com/gluster/glusterd2/tree/master/pkg/tools/)
+and save it in server nodes.
+
+Add alias in `~/.bashrc` as below,
+
+    alias gcurl='curl -H "$(python3 ~/glustercli-auth-header.py --secret-file=/var/lib/glusterd2/auth)"'
+
+**Note**: Change the path of script and auth file to match your setup.
+
+Thats all, use `gcurl` wherever `curl` is necessory. For example, to start a Volume
+
+    gcurl -XPOST http://localhost:24007/v1/volumes/gv1/start
+
+## Python example
+
+Install `jwt` library using `pip install jwt` or using `dnf install
+python-jwt`. Generate JWT token using,
+
+    import jwt
+    import requests
+
+    user = "glustercli"
+    secret_file = "/var/lib/glusterd2/auth"
+    secret = open(secret_file).read()
+
+    claims = {
+        "iss": "glustercli",
+        "iat": datetime.utcnow(),
+        "exp": datetime.utcnow() + timedelta(seconds=10)
+    }
+
+    token = jwt.encode(claims, secret, algorithm='HS256')
+    resp = requests.get("http://localhost:24007/v1/peers",
+                        headers={"Authorization": "bearer " + token}
+
+    peers = []
+    if resp.status_code == 200:
+        peers = json.loads(resp.content)
+
+
+## Using REST APIs from outside the Cluster nodes
+
+APIs for creating external users is not yet implemented. Temporarily
+copy the `auth` file from one of the server nodes and place it in a
+secured location.(Say `~/.glustercli/auth`)
+
+Run `glustercli` by specifying `--secret-file`. For example,
+
+    glustercli --secret-file=~/.glustercli/auth peer status
+
+Or use `curl` as below
+
+    curl -H "$(python3 ~/glustercli-auth-header.py --secret-file=~/.glustercli/auth)" \
+        http://gluster1.example.com:24007/v1/peers
+
diff --git a/pkg/tools/glustercli-auth-header.py b/pkg/tools/glustercli-auth-header.py
new file mode 100644
index 000000000..1a21b5032
--- /dev/null
+++ b/pkg/tools/glustercli-auth-header.py
@@ -0,0 +1,33 @@
+import sys
+from argparse import ArgumentParser
+from datetime import datetime, timedelta
+
+import jwt
+
+
+def jwt_token(user, secret):
+    claims = dict()
+    claims['iss'] = user
+    claims['iat'] = datetime.utcnow()
+    claims['exp'] = datetime.utcnow() + timedelta(seconds=10)
+
+    token = jwt.encode(claims, secret, algorithm='HS256')
+    return (b'Authorization: bearer ' + token).decode()
+
+
+if __name__ == "__main__":
+    parser = ArgumentParser()
+    parser.add_argument("--user", default="glustercli")
+    parser.add_argument("--secret-file", required=True)
+    args = parser.parse_args()
+
+    secret = ""
+    try:
+        with open(args.secret_file) as f:
+            secret = f.read()
+    except IOError as err:
+        sys.stderr.write("Unable to open secret file\n")
+        sys.stderr.write("Error: %s\n" % err)
+        sys.exit(1)
+
+    print(jwt_token(args.user, secret))