diff --git a/.github/workflows/client-matlab.yml b/.github/workflows/client-matlab.yml
index a751aeb..4d64cae 100644
--- a/.github/workflows/client-matlab.yml
+++ b/.github/workflows/client-matlab.yml
@@ -4,14 +4,14 @@ on:
   push:
     branches:
       - 'main'
-
+      
 jobs:
   test:
     runs-on: ubuntu-latest
 
     services:
       model:
-        image: linusseelinger/benchmark-analytic-banana
+        image: linusseelinger/benchmark-analytic-funnel:latest
         ports:
         - 4243:4243
 
@@ -24,4 +24,4 @@ jobs:
       - name: Run all tests
         uses: matlab-actions/run-tests@v1
         with:
-          source-folder: matlab; clients/matlab
\ No newline at end of file
+          source-folder: matlab; clients/matlab
diff --git a/clients/README.md b/clients/README.md
index 88f02c3..024544f 100644
--- a/clients/README.md
+++ b/clients/README.md
@@ -181,6 +181,83 @@ if (supports_apply_jacobian(url, name)) {
 
 [Full example sources here.](https://github.com/UM-Bridge/umbridge/tree/main/clients/R)
 
+## Matlab client
+
+The Matlab integration can be found in the [git repository](https://github.com/UM-Bridge/umbridge/tree/main/matlab).
+
+We use the Matlab function `addpath()` to add the specified folder.
+
+```
+umbridge_supported_models('http://localhost:4243‘)
+model = HTTPModel('http://localhost:4243‘, 'posterior');
+```
+
+`umbridge_supported_models()` gives a list of models that are supported by the current server. We set up a model by connecting to the URL and selecting for example the „posterior“ model. We obtain its input and output dimensions using the functions
+
+
+```
+model.get_input_sizes()
+model.get_output_sizes()
+```
+
+A model that expects an input consisting of a single 2D vector can be evaluated as follows.
+
+```
+model.evaluate([0, 10.0])
+model.evaluate([0, 10.0], struct('a', 3.9))
+```
+
+If the model accepts configuartion parameters, we can add those to the model evaluation. The config options accepted by a particular model can be found in the model’s documentation.
+
+Furthermore a model indicates wheather it supports further features. The following example evaluates the Jacobian of model output zero with respect to model input zero at the same input parameter as before. It then applies it to the additional vector given.
+
+```
+model.apply_jacobian([1.0,4.0], [0, 10.0], 0, 0)
+```
+
+[Full example sources here.](https://github.com/UM-Bridge/umbridge/blob/main/clients/matlab/matlabClient.m)
+
+## Julia client
+
+The Julia integration can be installed using Julia's builtin package manager Pkg
+
+```
+Pkg.add("UMBridge")
+```
+
+It can be used as usual via
+
+```
+using UMBridge
+```
+
+We set up a model by connection to the URL and selecting the "forward" model
+
+```
+model = UMBridge.HTTPModel("forward", "http://localhost:4242")
+print(UMBridge.model_input_sizes(model))
+print(UMBridge.model_output_sizes(model))
+```
+
+The functions UMBridge.model_input_sizes() and UMBridge.model_output_sizes() give the input and output dimension of a model, respectively. A model that expects an input consisting of a single 2D vector can be evaluated as follows
+
+```
+print(UMBridge.evaluate(model, [[0, 10]], Dict()))
+
+config = Dict("vtk_output" => true, "level" => 1);
+print(UMBridge.evaluate(model, [[0, 10]], config))
+```
+
+If the model accepts configuartion parameters, we can add those to the model evaluation. The config options accepted by a particular model can be found in the model's documentation.
+
+Models indicate whether they support further features, e.g. Jacobian or Hessian actions. The following example evaluates the Jacobian of model output zero with respect to model input zero at the same input parameter as before. It then applies it to the additional vector given.
+
+```
+ UMBridge.apply_jacobian(model, 0, 0, [[0, 10]], [1, 4])
+```
+
+[Full example sources here.](https://github.com/UM-Bridge/umbridge/blob/main/clients/julia/juliaClient.jl)
+
 ## MUQ client
 
 Within the [MIT Uncertainty Quantification library (MUQ)](https://mituq.bitbucket.io), there is a ModPiece available that allows embedding an HTTP model in MUQ's model graph framework.
diff --git a/clients/julia/juliaClient.jl b/clients/julia/juliaClient.jl
new file mode 100644
index 0000000..bc51770
--- /dev/null
+++ b/clients/julia/juliaClient.jl
@@ -0,0 +1,27 @@
+using UMBridge
+
+url = "http://localhost:4242"
+
+# Set up a model by connecting to URL and selecting the "forward" model
+model = UMBridge.HTTPModel("forward", url)
+
+# Print model input and output dimension
+print(UMBridge.model_input_sizes(model))
+print(UMBridge.model_output_sizes(model))
+
+param = [[0, 10]];
+
+# Simple model evaluation without config
+val = UMBridge.evaluate(model, param, Dict())
+print(val)
+
+# Model evaluation with configuration parameters
+config = Dict("vtk_output" => true, "level" => 1);
+print(UMBridge.evaluate(model, param, config))
+
+# If model supports Jacobian action,
+# apply Jacobian of output zero with respect to input zero to a vector
+if UMBridge.supports_apply_jacobian(model)
+    UMBridge.apply_jacobian(model, 0, 0, param, [1, 4])
+end
+
diff --git a/clients/matlab/matlabClient.m b/clients/matlab/matlabClient.m
index 21caf2a..58bf25e 100644
--- a/clients/matlab/matlabClient.m
+++ b/clients/matlab/matlabClient.m
@@ -1,5 +1,27 @@
+% use matlab function addpath('')
+
 uri = 'http://localhost:4243';
 
+% Print models supported by server
+umbridge_supported_models(uri)
+
+% Set up a model by connecting to URL and selecting the "posterior" model
 model = HTTPModel(uri,'posterior');
 
-httpValue = model.evaluate([1, 3])
+model.get_input_sizes()
+model.get_output_sizes()
+
+param=[0, 10.0];
+
+% Simple model evaluation without config
+model.evaluate(param)
+
+%Model evaluation with configuration parameters
+config = struct('a', 3.9);
+model.evaluate(param, config)
+
+% If model supports Jacobian action,
+% apply Jacobian of output zero with respect to input zero to a vector
+if model.supports_apply_jacobian()
+    model.apply_jacobian([1.0,4.0],param, 0, 0)
+end
diff --git a/julia/juliaClient.jl b/julia/juliaClient.jl
deleted file mode 100644
index 778ca99..0000000
--- a/julia/juliaClient.jl
+++ /dev/null
@@ -1,8 +0,0 @@
-using UMBridge
-
-url = "http://localhost:4242"
-
-val = UMBridge.evaluate(url, "forward" ,[[1,3]], Dict())
-
-print(val)
-
diff --git a/testing/clients/TestMatlab.m b/testing/clients/TestMatlab.m
index d717a7f..44db1fc 100644
--- a/testing/clients/TestMatlab.m
+++ b/testing/clients/TestMatlab.m
@@ -8,11 +8,22 @@ function testConnection(testCase)
             model = HTTPModel(uri,'posterior');
 
             httpValue = model.evaluate([1, 3])
-            exactValue = -11.194036030183454;
+            exactValue = -5.147502395904501;
             testCase.verifyEqual(httpValue, exactValue, 'RelTol', 1e-14)
 
         end
 
+        function testApplyJacobian(testCase)
+             uri = 'http://localhost:4243';
+
+            model = HTTPModel(uri,'posterior');
+
+            httpValue = model.apply_jacobian([1.0,4.0], [1,3], 0, 0)
+            exactValue = -3.370206919896928;
+            testCase.verifyEqual(httpValue, exactValue, 'RelTol', 1e-14)
+
+        end
+        
         function testBasicExample(testCase)
           matlabClient;;
         end