Hacker News Reader

nextimeon Self-hosted inference orchestrators compared: LocalAI, exo, GPUStack, vLLM

I wouldn't wrote as it, sorry, the post was an unintended generated post, wasn't supposed to post for real, until i was going to rewrite it by hand at least.

The meaning for that sentence is that litellm just route requests, it doesn't execute a model.

nextimeon Self-hosted inference orchestrators compared: LocalAI, exo, GPUStack, vLLM

not the whole blog ( but yes that post ), and yes it WAS autogenerated, it was supposed to be only a test but i did a gating wrong, apologise for it.

nextimeon Self-hosted inference orchestrators compared: LocalAI, exo, GPUStack, vLLM

sorry it was generated and was supposed to be a test and not to be posted. thanks for the llamafile suggestion!

nextimeon Self-hosted inference orchestrators compared: LocalAI, exo, GPUStack, vLLM

for whoever is reading and rightly say "it seems llm generated", you are right, sorry, wasn't supposed to really post it and it did by error, my bad for not supervision my test while i was doing it.

nextimeon Show HN: Aisbf, a self-hostable OpenAI-compatible AI proxy/router

still, you are commenting about a web page of a demo of an open source project instead to look at the project, you lame :)

and you are confusing a demo with a "commercial project". Look at the moon, not at the finger. Have a good life and if and when you will have any relevant comment or real critique, i will listen to you. PLONK.

nextimeon Show HN: Aisbf, a self-hostable OpenAI-compatible AI proxy/router

are you for real with this critique? out of a proxy/router packed full of functionalities that is also open source and self-hostable what you are looking at is the html of the demo service website that, by the way, has been written by gpt-5.5 ( so yes, an frontier llm did it )

You know when they say "when i point the moon you are looking at the finger"?

nextimeon Aisbf – AI Service Broker Framework (BETA) unified proxy and intelligent router

AISBF is now in BETA. It provides a single API endpoint that can route requests to OpenAI, Anthropic, Google, Ollama, and many other providers, with features like request splitting, semantic response caching, adaptive rate limiting, provider‑native caching, and OAuth2 integrations (Claude, Kilo, Codex, Qwen).

• Hosted demo (no setup): https://aisbf.cloud (https://aisbf.cloud/) • Self‑host: pip install aisbf • Source code: https://git.nexlab.net/nexlab/aisbf.git

We’d love feedback from the community—especially anyone juggling multiple LLM APIs who wants to simplify routing and reduce costs.

nextimeon Aisbf (AI Should Be Free) proxy 0.99.18 released

https://pypi.org/project/aisbf/

AISBF 0.99.18 released! In this version:

fix a LOT of bug full support without installing the cli and complete with Oauth2 initialization: claude subscriptions with oauth2 or api token codex cli subscription with oauth2 or api token qwen-code cli subscription with oauth2 or api token kilo is supported both with oauth2 or token kiro-cli full support any anthropic, gemini, openai, ollama compatible endpoints

nextimeon AIsbf (AI Should Be Free) 0.9.8 Released

https://pypi.org/project/aisbf/

AIsbf ( AI Should Be Frtee ) is a API proxy/router with intelligent ai driven router which exposes an openai compatible api to the clients making available to them in a unified interface different protocols and AI endpoint/services, offering various optimization aiming to make the costs of using LLMs more accessible to everyone.

It is multiuser, and can run from small setup or scale to big infrastructure.

In this last release: - support for cache on redis, sqlite, mysql, file - more context condensation method - native prompt caching and request caching support - faster and better semantic prompt based routing for autoselections - full support for Claude.ai subscribers with OAUTH2 - full support for Amazon Kiro-cli subscribers with OAUTH2 - full support for OpenAI codex subscribers OAUTH2 - full support for Kilo.ai subscribers using token or OAUTH2 - many bugfixes and new minor features

nextimeon Aisbf – an intelligent routing proxy for OpenAI compatible clients

AISBF - a personal AI proxy! Tired of API limits? Free accounts eating your tokens? OpenClaw needs snacks? This Python proxy handles OpenAI, Anthropic, Gemini, Ollama, and compatible endpoints with smart load balancing, rate limiting, and context-aware model selection with context condensation. Install with pip install aisbf - check it out at https://pypi.org/project/aisbf/

nextimeon Devuan Jessie 1.0.0 stable release candidate

The name "devuan" has been selected cause it's a merge of "debian" and "VUA" where "VUA" stand for "Veteran Unix Admins", the name of the group that started the fork.

It has to be pronounced as this: http://danex.nexlab.it/devuan.wav

The first release has been named "jessie" cause it's a 1:1 replacement of debian jessie, and for our luck jessie is also a name of a minor planet, so, it match our nomenclature and reflect that is a very close path to switch from debian jessie, as Devuan consider itself the "real" continuation of debian after wheezy.

nextimeon Devuan Jessie 1.0.0 stable release candidate

This is the right way to pronounce it:

http://danex.nexlab.it/devuan.wav

nextimeon Devuan Jessie 1.0.0 stable release candidate

http://danex.nexlab.it/devuan.wav

nextimeon Desktop privacy & security of web browsers on Linux part 1: concepts and theory

( anyway, for sure fix those issues in a compositor is easier and consume less resources than on X11, and then the wayland world is better from this point of view. But i don't think it's yet enough mature, and many WM/DE doesn't yet support it or support it only partially )

nextimeon Desktop privacy & security of web browsers on Linux part 1: concepts and theory

No, it doesn't.

Wayland have some advantages but also some disadvantages, and it isn't yet really mature.

Also, wayland, on the sandbox side, doesn't do anything special. As it's just a lot simple than X and does pretty anything except copy a buffer on the screen, it just delegate all IPC and rendering to the compositor, then, in the wayland world, the security issues of X11 are moved on the compositor instead to stay on the X server, but the issue are the same and the compositor must consider them.

nextimeon Desktop privacy & security of web browsers on Linux part 1: concepts and theory

yes, this isn't anything new, just an example setup of firejail and some sugar added.

nextimeon Desktop privacy & security of web browsers on Linux part 1: concepts and theory

not for sandboxing reasons, but i've used kvm + xpra for other things. It's usable, but of course the vm need a lot more resources

nextimeon Desktop privacy & security of web browsers on Linux part 1: concepts and theory

ok, you readed it and awhere of the -paranoid with xpra. Great. anyway, there is no way to obtain perfect security, this is just a little improvements as I explained.

nextimeon Desktop privacy & security of web browsers on Linux part 1: concepts and theory

Have you read the post?

Yes, X11 has no separation, but you can use multiple X11 isolated sessions, like i do in this config i'm explaining in this post.