Feature Request

[sharathdoes]

A small feature suggestion:

It'd be useful if the project supported an LLM service (through any provider, API key, or local model). that way, users could read a chapter and directly ask questions or summarize it. ( but I plan to fork and work on this actually now )

Thanks again for open-sourcing it.

[qymab]

@sharathdoes a docker image would be great.

[sharathdoes]

#4

tried my best @karpathy

[lsunay]

Take a look this fork https://github.com/lsunay/reader3-docker/tree/feature/docker-setup

[sharathdoes]

@lsunay great, works well, my #4 is slightly different, I'm using chatgroq as llm service, so user can use all available models with just one API KEY

[lsunay]

I use openrouter.ai and litellm to integrate ai proverders with local LLM's in one place. Same logic @sharathdoes

[sharathdoes]

ohh, thats great, I didn't completely see the code, pardon me. @lsunay

[lsunay]

To be clear it's not in the code. Litellm is a proxy server which connects API endpoints and serves from one point. You should look litellm specifications. By using litellm you can use seperate providers API in one point. So you don't have to use openrouter or grokchat services if you have seperate AI provider accounts. So simply litellm acts as grokchat or openrouter etc.
And thanks for your comments.

ohh, thats great, I didn't completely see the code, pardon me. @lsunay

[lee-b]

@sharathdoes a docker image would be great.

I wanted the same, so made this PR #15