v3.1.0 — Project Principles and Product Direction
v3.1.0
This is a docs-only release.
Summary
v3.1.0 adds the project principles and product direction for MLX Server Manager.
The goal is to clarify how future features should be evaluated as the project grows.
No app code was changed in this release.
Added
- Added
docs/product_direction.md - Added Project Principles to README
- Updated product direction and roadmap notes
- Updated known limitations
- Updated import/export documentation
- Updated onboarding guidance notes
- Updated task notes for v3.1.0
Project Principles
MLX Server Manager is guided by three principles:
- Preserve
mlx-lmruntime performance as the top priority. - Make
mlx-lmusable for users who are not comfortable with CLI workflows. - Adopt useful features from other local LLM tools when they do not conflict with
mlx-lmperformance, safety, or Direct Mode boundaries.
Product Direction
The project remains focused on helping users manage mlx_lm.server on Apple Silicon while keeping the runtime path thin and direct.
Direct Mode remains:
OpenAI-compatible client -> mlx_lm.server or adopted external server -> MLX model
Current Boundaries
MLX Server Manager currently does not:
- proxy inference requests
- act as a chat UI
- perform multi-backend routing
- rewrite requests
- automatically take ownership of external processes
- automatically delete models
- silently download models
- start servers after import or download
Model Download Position
Model download is not implemented in the current release.
It may be considered in the future if it:
- does not reduce
mlx-lmruntime performance - does not interfere with managed server operation
- keeps clear safety and privacy boundaries
- does not silently download files
- does not automatically start servers
- does not delete models
- clearly shows target location, progress, and failure state
Current Binary Asset
The current downloadable app binary remains the v3.0.0 unsigned build:
MLXServerManager-v3.0.0-unsigned.zip
This v3.1.0 release does not include a new binary asset.