Back to Directory
research
Offload Tasks to LM Studio Models
Equips agents to search for and offload tasks to local models in LM Studio. Use when: (1) Offloading simple tasks to free local models (summarization, extraction, classification, rewriting, first-pass code review, brainstorming), (2) Tasks need specialized model capabilities (vision models for images, smaller models for quick tasks, larger models for complex reasoning), (3) Saving paid API tokens by using local models when quality is sufficient, (4) Tasks require local-only processing or privacy. Requires LM Studio installed with lms CLI and server running.
Installation
Run this in your terminal or add to your configuration:
# Clone into your skills directory
git clone ...About this Skill
- Added Node.js helper script `scripts/lmstudio-api.mjs` for robust LM Studio API calls.
- Updated documentation to recommend using the helper script for making API requests (with fallback to curl).
- Clarified that no OpenClaw configuration for API URL is needed; the default LM Studio server is sufficient.
- Noted Node.js as a new prerequisite for using the script.