Power Platform Community Forum Thread Details

I have this flow which generates a PDF with text and links which comes from a lookup column in Lists. The Flow generates perfect PDF EXCEPT the links are not clickable. Well, the browser tells me that they try to open file:///C:/MetaTempData/ProcessFiles/[number]/input.html, i.e. the links are not generated correctly in the Flow. The crazy thing is that I created and tested the flow afew months ago and then the links worked fine.

I have attached screenshot of my Flow and the CodeView of the main scope. The links are public links: https://learnit.itu.dk/local/coursebase/view.php?ciid=2087&view=public and https://cloud.timeedit.net/itu/web/public/ri11064y69X056Q1ZpQ10Z60Z05c85QQ5YX3gY15xY1Q391XY46Zg95g0564Y29Q0.html

Do you have any ideas? Thanks in advance

Data, Automation ...

RoleDescriptionFo...

Categories:

Using flows

Hi @Rikke1,

This type of issue usually happens when Power Automate’s HTML --> PDF renderer changes how it interprets links inside the HTML you generate. Even though nothing changed in your Flow, the backend rendering engine can update, and suddenly URLs stop being treated as absolute hyperlinks and instead get interpreted as local file paths such as:

file:///C:/MetaTempData/ProcessFiles/.../input.html

This is exactly what happens when a link inside the HTML is missing a proper href="https://..." or when the renderer fails to recognize the URL as an external HTTP link.

Here are the most reliable fixes:

1. Make links explicitly absolute
Ensure the URL inside your HTML has the full protocol and domain, e.g.:
HTML<a href="https://learnit.itu.dk/local/coursebase/view.php?ciid=2087&view=public"> Course Link</a>Show more lines
Even if your content looks like a full URL, sometimes the HTML generator escapes characters or converts the href into a relative link.

2. Disable HTML encoding before PDF creation
If you are feeding dynamic text from a Lists column, Power Automate sometimes encodes the URL like:
https%3A%2F%2Fcloud.timeedit.net%2F...

This causes the renderer to treat it as a literal string.
Fix:
Use “Compose” or “Replace” to decode %3A --> :, %2F --> /, or use:
decodeUriComponent(<yourLink>)

3. Ensure your HTML contains real anchor tags
If you're inserting raw URLs like:
https://cloud.timeedit.net/...

PDF conversion will often convert them to plain text → no link.
Wrap them manually:
<a href="@{items('Apply_to_each')?['URLColumn']}">Open link</a>Show more lines

4. If you're using Convert HTML to PDF (OneDrive for Business)
That action is known to occasionally regress and treat links as local resources.
More stable workaround:

Use the Word Template --> PDF approach
(Word maintains link metadata far more reliably)
Or use Encodian / Muhimbi / Plumsail PDF connectors if available

5. Compare working vs broken HTML
When the issue comes and goes, copy the final HTML content of both versions from “Peek code”.
You’ll likely see something like:

Missing "https://" prefix
HTML‑encoded characters inside href
Link placed inside text but not inside an <a> tag

Fixing these ensures stability even if backend behavior changes again.

✅ If this answer helped resolve your issue, please mark it as Accepted so it can help others with the same problem.
👍 Feel free to Like the post if you found it useful.