AI Kits

Anthropic Cookbook

OSS

Official patterns library from Anthropic.

— First place to look when designing an agent prompt.

DSPy

OSS

Prompts as programs you compile.

— Compelling research direction; not quite production-ready as of 2026.

LangChain

OSS

Useful primitives, controversial abstractions.

— We took inspiration, not dependency. Worth reading even if you don't adopt.

LlamaIndex

OSS

Focused on RAG and document ingestion.

— Better defaults than LangChain if your problem is really retrieval.

Hugging Face Datasets

OSS

Hugging Face's datasets library + hub.

— Essential infrastructure for fine-tuning; also great for eval datasets.

Inspect AI

OSS

UK AISI eval framework.

— What we'd use for before-you-ship LLM tests on any serious system.

OpenTelemetry GenAI

OSS

OpenTelemetry semantic conventions for LLM traces.

— The emerging standard. We wired the VORLUX orchestrator against this spec.

promptfoo

OSS

Config-driven eval harness.

— Pairs well with CI. YAML-first approach that scales past one person.

LM Studio

freemium

GUI alternative to Ollama if you prefer a window to a terminal.

— Identical capability, friendlier onboarding. Good for client demos.

Ollama

OSS

Local LLMs, one command.

— The backbone of VORLUX's local-first stance. Everyone should have this installed.

vLLM

OSS

High-throughput inference server.

— When you outgrow Ollama and need real QPS — usually past 10 concurrent users.