Lemonade v10.8 automates local AI memory and optional cloud use
Lemonade v10.8 is an update for running AI models on a local computer with less manual setup. The release had 20 contributors and was completed in 7 days. It can now unload idle models by itself and shrink the KV-cache to free memory while the system is running.
Models that should stay ready can be pinned so they are not removed from memory. Lemonade also chooses the automatically based on available memory and the model design, instead of making the user tune it by hand. When a larger model is needed, work can be sent to providers such as Fireworks, OpenRouter, Together, and OpenAI.
This sits beside local models as an option, while local running remains the default. LMX-Omni adds more controls, such as image size and steps, and can pull or share custom Omni models through . Local models can also be called as MCP tools, which makes them easier for other apps or agents to connect to.
Key points
- Idle local models can be unloaded automatically to recover memory.
- Pinned models can stay ready instead of being removed from memory.
- can be chosen automatically from the machine’s memory and model design.
- can send requests to providers when a bigger model is needed.
- Local models can be exposed as MCP tools for apps and agents.