The video, presented by Jonathan Huberman on his YouTube channel (youtube.com/@JonathanHuberman), demonstrates how to automatically turn AI-generated conversations from NotebookLM, or any podcast audio, into video podcast episodes using JoggAI. The main problem addressed is the time-consuming and multi-step manual process often required to produce video podcasts. JoggAI offers an automated, two-step solution that can be completed in less than 10 minutes. Here’s a breakdown of the processes shown:
- Generate Audio Dialogue with NotebookLM:
- Jonathan first creates a NotebookLM notebook on “Vibe Coding” using a YouTube video and a Wikipedia page as sources.
- He then uses NotebookLM’s “Audio Overview” feature to generate a realistic 19-minute audio dialogue based on these sources.
- This audio is downloaded as a WAV file.
- Turn Audio into Video with JoggAI (Keeping Original Voices):
- He navigates to the JoggAI website (link available in the video description) and logs in.
- Within JoggAI’s “Video Podcast” section, he selects “Upload Audio” and uploads the NotebookLM-generated WAV file.
- JoggAI quickly processes the audio and generates a script.
- Jonathan then chooses a two-person avatar pair (matching the gender of the original NotebookLM speakers) and a visual theme (e.g., “Mountain Lodge”). Crucially, he notes that when uploading audio, JoggAI retains the original voices, so no new voices need to be selected within JoggAI.
- The platform then renders the video with the chosen avatars speaking the NotebookLM dialogue, synced automatically.
- Swap Out NotebookLM Voices (Optional):
- For those who want unique voices, JoggAI allows users to swap out the original NotebookLM voices.
- Instead of “Upload Audio,” the “Upload Script” option is used. After uploading the audio file, JoggAI extracts the transcript, allowing for editing.
- Users can then select an avatar pair and explicitly choose different AI voices from JoggAI’s library for each speaker.
- Podcast Video From Any Podcast / Any Source Type:
- Jonathan highlights that JoggAI is not limited to NotebookLM audio. It can generate video podcasts from _any_podcast audio (e.g., from Jellypod, another AI podcasting platform he uses).
- Furthermore, JoggAI’s “Generate Script” option can create a video podcast from diverse source types such as website URLs, YouTube video links (which are then used to create an original script, not a word-for-word transcript), PDF documents, or even long-form text. Users can specify the desired video and turn durations, and JoggAI will summarize longer content.
**Conclusion:**JoggAI provides a powerful and automated solution for turning audio dialogues into engaging video podcasts with talking avatars, saving significant time and effort. While early versions of the direct audio conversion might have minor sync issues, the platform is continuously improving. It’s a valuable tool for expanding podcast reach on video platforms, whether using original or AI-generated voices, and from various content sources. Jonathan also provides links in the video description to additional free resources, including a complete ChatGPT course and an AI tools ecosystem guide.