Skip to main content

Hi everyone!
I’m currently trying to build a Zap that transcribes audio files using AssemblyAI, but I’m running into two blocking issues and could really use your help:

❶ When using Download URL:

I get this error from Zapier:

"Failed to create a transcript in AssemblyAI. The app returned: ‘No valid API key was provided in the request.’”
→ My credit card is registered and the API key is correctly set in Zapier. Not sure why it’s not recognized.

❷ When using File instead of URL:

AssemblyAI returns a completely unrelated transcript (seems like a sample).
It looks like it’s transcribing this instead:

https://assembly.ai/wildfires.mp3

But I never provided that file — I'm trying to send my own M4A from Google Drive.

Has anyone experienced this?
Any idea how to fix these two problems?

I’d really appreciate any advice!

Hey ​@Tomoya ,

 

Please share a screenshot on how did you setup your Zap.


Hi ​@Tomoya,

 

If you're having trouble connecting AssemblyAI with Zapier, start by checking that your API key is correct, has the proper permissions, and is connected correctly in Zapier. For transcript issues, ensure the audio file URL is directly accessible, in a supported format, and mapped correctly in your Zap. Testing with real audio data can help identify problems, and if issues persist, contacting Zapier Support is recommended for further assistance.

I hope this helps. Please let me know if you have any more questions or issues.


Hi there, ​@Tomoya 👋

Were you able to get this sorted? 

If not, for the “No valid API key was provided in the request.” error—is the API Key from a paid, cloud-hosted account? If it is, it might be worth regenerating the API Key in AssemblyAI, and reconnecting the AssemblyAI account connection to see if that then allows the Transcribe action to run successfully.

With the sample transcript you’re seeing, that appears to be the expected behavior when testing in the Zap editor:

ae4af75c25888329752a47b9b19dc649.png

Once the Zap is turned on and running live, it should process the actual files it’s provided rather than the sample transcripts.

Keep us posted on how you’re getting on with this, want to make sure everything’s sorted. 🙂