 | using qwen 3 VL for the llm and the vision (really good for recognize popular characters and even recognize their appearances) using SerpApi for the web search the tts is using omnivoice tts (support 600+ languages) that i make a custom api that i recently open source it, get it here: https://github.com/aziib/omnivoice-tts-api my ai waifu project stil in work in progress, i just hope there is free web search api, SerpApi has some search limit usage per month. [link] [comments] |