Pin the embedding model to the CPU so ZeroGPU stops intercepting it

Repair the edits the web editor mangled, and the paths they broke

Update src/llm/langchain_llm.py

Update src/llm/langchain_llm.py

Update src/llm/langchain_llm.py

Update src/core/engine.py

Create langchain_llm.py

Kalim Space deploy (e803a10)