Chat2B - one client-side chat UI for any OpenAl-compatible API
Hey everyone — Mateus here, from Brazil.
I build Chat2B, a client-side chat interface for the major language models. The idea is one simple, universal chat UI: it talks to anything that speaks the OpenAI API standard — Gemini, Groq, xAI, custom endpoints — and to local models behind Ollama, llama.cpp, transformers, or anything else exposing an OpenAI-compatible /v1.
Two paths, on purpose. With a public API there's no backend at all: open the page, paste your key, chat — requests go straight from your browser to the provider. With a local model you run your own backend on your own machine, and Chat2B connects to it like any other endpoint. Either way, keys and conversations stay with you, not on someone else's server.
What's working: keeping it genuinely provider-agnostic — cloud or local is the same UI, just a different base URL.
One thing I need: honest feedback from people who've used client-side chat UIs — what actually makes you pick one over the official apps?