diff --git a/06_coding_patterns.md b/06_coding_patterns.md
index 6b0752c..d344ff7 100644
--- a/06_coding_patterns.md
+++ b/06_coding_patterns.md
@@ -685,10 +685,10 @@ An alternative is to let the solver run in a separate process and communicate
 with it using a pipe. This approach allows the solver to be interrupted at any
 time, enabling the application to react immediately. Python's multiprocessing
 module provides reasonably simple ways to achieve this.
-[This example](https://github.com/d-krupke/cpsat-primer/blob/main//examples/embedding_cpsat/) showcases such an approach. However,
-for scaling this approach up, you will actually have to build a task queues
-where the solver is run by workers. Using multiprocessing can still be useful
-for the worker to remain responsive for stop signals while the solver is
-running.
+[This example](https://github.com/d-krupke/cpsat-primer/blob/main//examples/embedding_cpsat/)
+showcases such an approach. However, for scaling this approach up, you will
+actually have to build a task queues where the solver is run by workers. Using
+multiprocessing can still be useful for the worker to remain responsive for stop
+signals while the solver is running.
 
 ---
diff --git a/README.md b/README.md
index 31218cd..cfdf9a0 100644
--- a/README.md
+++ b/README.md
@@ -3119,11 +3119,11 @@ An alternative is to let the solver run in a separate process and communicate
 with it using a pipe. This approach allows the solver to be interrupted at any
 time, enabling the application to react immediately. Python's multiprocessing
 module provides reasonably simple ways to achieve this.
-[This example](./examples/embedding_cpsat/) showcases such an approach. However,
-for scaling this approach up, you will actually have to build a task queues
-where the solver is run by workers. Using multiprocessing can still be useful
-for the worker to remain responsive for stop signals while the solver is
-running.
+[This example](https://github.com/d-krupke/cpsat-primer/blob/main//examples/embedding_cpsat/)
+showcases such an approach. However, for scaling this approach up, you will
+actually have to build a task queues where the solver is run by workers. Using
+multiprocessing can still be useful for the worker to remain responsive for stop
+signals while the solver is running.
 
 ---