Ettore Di Giacinto

@LocalAI_API 4.0 is out and its crazy - New UI with React (huh!) - Canvas mode in chat - Agentic orchestration, Memory (Hybrid search), Skills management integrated - WebRTC for realtime (@jichiep ) - New backends (ace-step.cpp, faster-qwen3-tts) And this is just the tip of it👇

wow, that's a weird security profile. wild times ahead

@evilsocket @LocalAI_API no worries, I'm happy to support. I've seen Cake from the day 0 and was always intrigued by it and I'd be happy to give it more visibility as much as I can. Also @lu_zero_ is a rust fun so I'm pretty sure he would be interested to help too

@mudler_it @LocalAI_API I appreciate it, thanks! Apologies for previous interactions, I have been multiplexing between too many projects at once :D

Generated on my laptop. I'm preparing Cake to be the fastest multimodal inference engine for text, image and voice, running on consumer hardware either in single node or clustered. Android and iOS included.

@evilsocket @LocalAI_API Fantastic!! I can also help there! Id be happy to shout it out in our release and have dedicated section in our docs ❤️

@mudler_it @LocalAI_API You know what? I think LocalAI might add all the fancy UI/UX that I really don't feel like adding to Cake :D Once I take it to 1.0.0 I would be very happy to integrate

@evilsocket @LocalAI_API Upstream* :)

I don't want to sound rude but did you check closer? This is not really true as we acknowledge all inferencing engines and are really visible to the user what they are using, also we reference all engines used by LocalAI. Damn, we also contributed to downstream inferencing engines..

@evilsocket @LocalAI_API See it like a way to give cake to more users :) LocalAI while it seems adding layers, actually it smartly downloads the backend needed for the models, anyhow - like you prefer I thought it could be nice for open source project to help each other!

Thanks Ettore, with all due respect for the awesome project, LocalAI is the exact opposite of what I want for Cake. Cake is 100% written in Rust, self contained, little to no dependencies, with custom fused CUDA, Metal and soon AMD kernels. The objective is to make it faster than any of the python / go / cpp backends the project uses, easier and cheaper to deploy, etc.

@UnslothAI Fantastic! so curious, gonna play with it for sure

Introducing Unsloth Studio ✨ A new open-source web UI to train and run LLMs. • Run models locally on Mac, Windows, Linux • Train 500+ models 2x faster with 70% less VRAM • Supports GGUF, vision, audio, embedding models • Auto-create datasets from PDF, CSV, DOCX • Self-healing tool calling and code execution • Compare models side by side + export to GGUF GitHub: github.com/unslothai/unsl… Blog and Guide: unsloth.ai/docs/new/studio Available now on Hugging Face, NVIDIA, Docker and Colab.

Can someone explain to me why the hell an agent's sandbox should have vim or nano?

It's hard to get for some, but this is the harsh reality now. There is no tech stack anymore that provides value by per-se. It doesn't only need to fix real problems, its the whole marketing/sell era is ahead of us.

This is so cool. You should check out latest @jichiep contributions, are really improving @LocalAI_API usage by a tons

@ibuildthecloud my english sucks roflol. I meant like, someone that doesn't work in the tech industry by itself (like, devs hired directly by assurance companies to work on internal processes, etc)

@mudler_it What is a out-of-tech dev? Like a smb dev?

I really have so little tolerance for enterprise AI. I talk to small to medium size businesses and it's so refreshing. They have obvious problems that AI can solve. Enterprise just wants solutions to control the thing they are afraid of that doesn't even exist.

Working on #MocaccinoOS installer to match the new background image. #Linux #Artwork

@mil0theminer Hype train

Why do people talk about OpenClaw like it’s a model and not just a harness? It’s not doing anything revolutionary

Created the @LocalAI_API community on X x.com/i/communities/…

@jestermolecule @LocalAI_API @jichiep yes, I've set Github sponsorship 👍

@mudler_it @LocalAI_API @jichiep sure thing man! could i just donate fees directly to your GitHub to support?

@LocalAI_API 4.0 is out and its crazy - New UI with React (huh!) - Canvas mode in chat - Agentic orchestration, Memory (Hybrid search), Skills management integrated - WebRTC for realtime (@jichiep ) - New backends (ace-step.cpp, faster-qwen3-tts) And this is just the tip of it👇

@LocalAI_API @jichiep @FishAudio @andimarafioti

@LocalAI_API @jichiep @ACEStudio_en

@jestermolecule @LocalAI_API @jichiep yes of course!

@mudler_it @LocalAI_API @jichiep could i make a donation to ur github for this?