Roadmap
From private alpha to the first public release.
LM Nexus already works as a private alpha. A few important pieces still need to land before a public build. Here is what works today, what is being built now, and what comes later.
Current
What works today
The current private alpha already covers local model management, runtimes, providers, agents, and a broad project workspace.
- download and run local models with llama.cpp
- connect directly to OpenAI-compatible runtime endpoints
- use local, self-hosted, custom, and connected providers
- chat and agent workflows
- Files, Notes, Knowledge, Terminal, and Git
- modules, packages, and Marketplace surfaces
Preview
What we're building now
Most current work is about making longer agent tasks survive interruptions, keep useful memory, and recover cleanly.
- Durable Work Runtime
- Memory and Context Builder
- Model Router
- retries, recovery, and supervision
- human handoffs and deeper delegation
- cleaner settings, website, and docs
Planned
Before public release
After the current feature work comes licensing, platform testing, release packaging, and the final public documentation.
- finish the move to AGPL-3.0-only
- public installers and release packages
- release notes and changelog tooling
- complete public getting-started docs
- final platform and support validation
- open the module ecosystem more broadly
Looking ahead
The order can still change.
This is a direction, not a promise that every item will ship in this exact order. Dates and release details will get more concrete as the remaining product, platform, and licensing work settles.