Check out my workflow: Writing in Notion, exporting via API to PlayHT, importing to Descript.

Been trying to streamline my podcast production for the marketing show I run, and I think I've finally nailed a workflow that saves me a ton of time. It uses three core tools: Notion, PlayHT, and Descript. The goal was to keep the writing and editing in a comfortable place, automate the voiceover creation, and then do final polish easily.

Here’s the basic flow:
* I draft all my scripts in Notion. I have a database with fields for episode title, key points, and the full script. The structure just works for me.
* Once a script is ready, I use Notion's API to push the text directly to PlayHT. I have it set up to use a specific voice clone I trained for consistency.
* PlayHT generates the audio file and drops it into a designated cloud folder.
* That folder is watched by Descript, which imports the audio automatically. From there, I can edit out mistakes, add music, and publish.

The biggest wins for me:
* **No more copying and pasting** text between apps. The API handoff is seamless.
* **Voice consistency** across episodes without needing to record myself every time.
* Editing in Descript is still way faster for me than any DAW when it comes to cutting filler words and small mistakes from the AI audio.

I did run into a couple of snags initially. The formatting from Notion (like bullet points) sometimes needed cleaning up for PlayHT to read naturally. I ended up adding a simple formatting step in my automation (Zapier, but you could use Make or even a custom script). Also, getting the right pauses and emphasis in the PlayHT script took a bit of trial and error with their SSML tags.

Overall, this has cut my production time per episode by about 60%. Curious if anyone else is bridging these tools or has found better ways to handle the text-to-speech step. The less time spent on mechanics, the more time for content and promotion!

— alex