AI

AI/LLM-powered features — the BYOK AI chat, and any agentic or generative features layered onto Localz or FullStackFusions.

Why we need this / what value this brings

Gives Localz/FullStackFusions a way to layer generative or agentic features on top of existing product without betting the whole product on one AI vendor.

When to use this

When a feature needs natural-language understanding/generation that hardcoded logic can't reasonably do — chat, summarization, semantic search, matching.

How to use or implement this

Start with LLMIntegration (a single provider, plain request/response) before reaching for RAG or agents — add complexity only once the simple version is proven insufficient.

Subtopics

• AgentMemory ShortTerm LongTerm Summarization
• AgenticWorkflows
• ContextWindowManagement ChunkingStitching
• LLMGateway EnterpriseRouting
• LLMIntegration
• PromptEngineering
• RAG
• VectorStores

Research questions

• Which of these subtopics does the BYOK AI chat (OpenAI/Anthropic/Groq/Ollama) actually need first?
• Where would Localz itself use AI — e.g. provider/service matching, auto-generated listing copy, support chat?

---

Empty folder — drop notes, links, and findings here as you research.