Hi! Yes, the premium voices are Kokoro. I’m only exposing the English voices right now because the rest of the pipeline around them is English-first and custom, especially pronunciation/G2P, QA, and timestamp awareness. I’d like to expand that over time, but I don’t want to overpromise multilingual support before the surrounding stack is ready. So I'm taking it one language at a time based on demand and feedback.
AI summaries are currently generated remote, not local. Those currently leverage gpt-4o-mini. TTS and OCR are on-device and summarization is the cloud-backed feature.
I like this approach and that it's so flexible and approachable. My nit for the site would be to explain it a bit better - it's a bit hard to grok with everything going on until I went to the repo. There's also opportunity for non-tech users here just using the agent skill but I'm not sure they'll understand the use if they read through the site.
Agree. I keep effort max on Claude and xhigh on GPT for all tasks and keep tasks as scoped units of work instead of boil the ocean type prompts. It is hard to measure but ultimately the tasks are getting completed and I'm validating so I consider it "working as expected".
AI summaries are currently generated remote, not local. Those currently leverage gpt-4o-mini. TTS and OCR are on-device and summarization is the cloud-backed feature.