Local AI workflows are powerful, but they are often scattered. One tool runs models. Another tool manages chat. A terminal starts a server. Notes live somewhere else. Files, agents, logs, provider dashboards, and scripts all compete for attention.
LM Nexus is an attempt to bring those layers together without hiding the underlying runtime from people who need to understand it.
Simple versus controllable should not be a tradeoff
Beginners should be able to choose a model, click Run, and start working. Power users should be able to inspect runtime settings, copy endpoints, view logs, and connect directly to the server when they need low-level control.
LM Nexus is designed to support both paths. The goal is click-and-run local AI for beginners, raw runtime control for power users, and a modular workspace for everything around it.
Direct runtime/server access matters
LM Nexus is built around direct raw-server access with no additional Nexus inference layer required.
When users connect directly to a llama.cpp server started by LM Nexus, inference traffic does not pass through a Nexus inference proxy. In that mode, Nexus acts as an orchestrator/manager rather than an inference middleware layer.
That matters for debugging, transparency, compatibility, and power-user workflows where the raw endpoint is part of the system design.
curl http://127.0.0.1:PORT/v1/chat/completionsLocal-first does not mean local-only
Local-first is a product stance, not a provider lock-in. LM Nexus is designed for local models, self-hosted runtimes, custom OpenAI-compatible endpoints, and connected/API providers.
Some work belongs fully local. Some teams have self-hosted inference. Some workflows need a connected provider. LM Nexus is designed to make those choices visible and configurable from one control center.
The workspace layer matters
Models are only part of AI work. The surrounding workspace includes agents, files, notes, knowledge, terminals, logs, dashboards, settings, and future add-ins.
LM Nexus treats those as first-class pieces of the product rather than afterthoughts. Chat can connect to agents. Logs can become context. Notes can become knowledge. Modules can expose capabilities for optional cross-module actions.
Current status
LM Nexus is currently in pre-public/private-alpha preparation. It is not publicly downloadable yet, and this site intentionally does not invent release links.
Public builds are planned after the private alpha stabilizes. Until then, the best way to use this site is to understand the product direction, review the docs, and follow development as public links become available.