diff --git a/README.Rmd b/README.Rmd
index 9c46849..2540d30 100644
--- a/README.Rmd
+++ b/README.Rmd
@@ -162,7 +162,7 @@ The above examples present a simple all-in-one workflow, which is concise but ca
 It is primarity aimed at workflows which prioritise updating data: every time the script is run, data is redownloaded, unless cached via the `dest_dir` parameter.
 
 In other situations, the priority might be to keep track of individual source files as they are downloaded from the data provider, or checking for changes at the data provider and keeping track of individual URLs from which the data was downloaded.
-For these situations, a workflow composed of lower-level functions is available, offering finer control of the steps. See the [workflow](/articles/workflow.html) vignette (`vignette("workflow", package = "statnipokladna")`).
+For these situations, a workflow composed of lower-level functions is available, offering finer control of the steps. See the [workflow](https://petrbouchal.xyz/statnipokladna/articles/workflow.html) vignette (`vignette("workflow", package = "statnipokladna")`).
 
 
 
diff --git a/README.md b/README.md
index 3639b4d..d4db4e5 100644
--- a/README.md
+++ b/README.md
@@ -267,8 +267,9 @@ source files as they are downloaded from the data provider, or checking
 for changes at the data provider and keeping track of individual URLs
 from which the data was downloaded. For these situations, a workflow
 composed of lower-level functions is available, offering finer control
-of the steps. See the [workflow](/articles/workflow.html) vignette
-(`vignette("workflow", package = "statnipokladna")`).
+of the steps. See the
+[workflow](https://petrbouchal.xyz/statnipokladna/articles/workflow.html)
+vignette (`vignette("workflow", package = "statnipokladna")`).
 
 ## Background information