Anything AI
- 7 Devlogs
- 5 Total hours
Uses multiple models on Hugging Face and another intent routing model, so that your request is handled by the best open source LLM on the job. REQUIRES HUGGING FACE READ ACCESS TOKEN
Ship #1
Pending review
Damn man…. so much for adding chat history and long term memory….
Today I spent 1 and a half hours adding them, the memory is stored in Anything AI\memory and chats are stored in Anything AI\chats. I also switched the sorter AI to deepseek-ai/DeepSeek-R1-Distill-Llama-70B coz llama was hallucinating a LOT, ill be releasing another beta soon…..
I’m regretting making a TUI instead of a GUI.
I made it so because i wanted to save time and focus on the important stuff, and cause i was lazy, but it turned out to be a mistake.
Today I spent nearly an hour trying to fix terminal formatting issues like shown in the attached screenshot (sorry for low resolution) , it got so bad that I just ended making my own input() function instead. Truly an experience, anyways, Anything has entered beta! The full proper official ship will come soon!
Now I have created the logo for Anything AI, so it will be definitely out soon! I hope this project gets high ratings!
YESS, attach file works! It currently supports .txt, .csv, .docx, .pptx, .xlsx and code files like .py, .kt, .java, .cpp and, obviously image files, the ones supported by PIL :). Anything AI will be finished soon!
Today I’m working on adding more modalities to my AI, more specifically, giving it the ability to read text, documents and code files alongside the already implemented image file. To do this I’m migrating from pick_image_file() to just pick_file()
As of now, the pick_file() function is done and i need to add functions for parsing the returned filepath to data the AI can read.
I also ran out of free credits on Hugging Face so i created an extra account. :(
OK, so terminal commands work! I also fixed the jank bug that caused issues with displaying text after addingf an image to a prompt. Things seem to be going well! I’m gonna make myself a bowl of maggi now >. <
Ok so, I know this is a project I started before Stardance on May 25 2026 and have worked for hours on this, and its almost ship ready, I’ve added everything I want like multimodal input, file generation, etc.
As of now, I’m giving it the ability to run terminal commands and do some bugfixing before shipping it as an .exe
