ls-local-lens

What problem does it solve?

This Skill prevents context bloat by compressing, extracting, and summarizing large content using locally loaded LM Studio models instead of sending raw text to a remote LLM.

Core Features & Use Cases

• Context compression for large inputs: Turn 15,000+ tokens of raw material into short, structured summaries suitable for planning and downstream reasoning.
• Local extraction pipelines: Extract factual fields from long logs, API responses, and documents with deterministic, schema-like prompts.
• On-demand model routing: Automatically route tasks to the most appropriate local model (summarize, code analysis, RAG-style Q&A, or reasoning) while managing VRAM safety with a switcher.
• Vision-aware description workflows: Describe screenshots using a VLM model loaded in LM Studio to avoid wasting tokens on image data in the main context.

Quick Start

Ask your agent to "compress this context before answering" and route the request through Local Lens using the local-lens triggers such as "summarize this" or "offload to local model".