WHAT IT DOES

What does it do?

Galahad stores the results of a model reading text and provides them again when needed. It also stores the documents themselves and finds sections relevant to a question so the model reads only what it needs. It has a C++ core and runs inside llama.cpp, vLLM, and SGLang.

KEY FEATURES

Key features

• Stores and reuses the results of a model reading text
• Finds document sections relevant to a question
• Runs inside llama.cpp, vLLM, and SGLang