3 ms·
Show HN: Clai – OpenAI models brought to the terminal
Supports text and image generation. For text generation you can query, chat or enter a glob + a query. For any of the commands xargs-like argument replacement can be used.
It's written using only go's standard library (plus go experimental, within my boilerplate package)
- mehmet_mhy 3y agoThis looks awesome. I like how clean it looks. I made something similar called Cha (https://github.com/MehmetMHY/cha https://github.com/MehmetMHY/cha) that supports image generation, answer search, and web scrapping.
- baalimago 3y agoVery cool! The way to swap between models and 'live configure' is neat, also the ascii picture previews. I think the idea of a cli text/photo generator is so simple I think that anyone who is technically adept enough to appreciate it are also able to write a version of their own within a short amount of time.
- jshreder 3y agoThis looks great -- I've been using https://github.com/ibigio/shell-ai https://github.com/ibigio/shell-ai (aliased to `q` ) but this looks even more apt for my use case. I use TypingMind for any real conversation with LLMs, but for quick answers in terminal, these kinds of tools are super useful. I love the `ask` and `rask` shortcuts! Would love support for Claude APIs :)
- baalimago 3y agoThanks, I'm glad you like it! I plan on adding support for Claude whenever it available in Finland.
- jshreder 3y agoAwesome! I might open a PR for Gemini 1.5 in the meantime too, its huge context is awesome for dumping files into it
- baalimago 3y agoHeard from a friend that I can just VPN to USA and setup claude that way. With registration done, the account remains active (apparently). So my easter project will be to scale up clai to support multiple ai-model vendors with claude and also add midjourney/stable diffusion for photo generation. Once that's setup, I'd highly appreciate your addition for gemini 1.5!