Make it a hybrid product. Let the end user also mark elements such as video, slides, original audio, and transcription. Especially of aligning speech-to-text to the generated audio, video and slides. Allowing the user to scrub through the timeline and mark what they believe is important.
This allows the user to be at the center of the product, yet scale their use case and manual input as needed with generative AI. In addition, this provides additional context for the AI to produce tailored output based on the user's input.
My comment is not about LLM provider lock-in. It's about merging a traditional note-taking app with an LLM augmented approach. It's up to you if you want your product to stand out.
So I'm tired of existing AI notetakers. Building my own. Key ideas:
Your device only - All recordings stored on your device.
Your choice of LLM - use any preferred LLM.
Your existing workflow - easy connect to existing tools like Notion, Slack, etc..
Make it a hybrid product. Let the end user also mark elements such as video, slides, original audio, and transcription. Especially of aligning speech-to-text to the generated audio, video and slides. Allowing the user to scrub through the timeline and mark what they believe is important.
This allows the user to be at the center of the product, yet scale their use case and manual input as needed with generative AI. In addition, this provides additional context for the AI to produce tailored output based on the user's input.