Some categories are for people who build AI features into their own products, not people who use a finished tool. If that's you, here is how the pieces fit together. If you're new to the terms, start with the glossary.
The model layer
LLM Platforms
LLM Platforms lists large language model providers, hosting services and APIs. This is where you get access to a model. Compare quality for your task, price per use, speed, context size, data handling and any rate limits.
Datasets
Datasets contains training and evaluation data and tools for labeling it. Look carefully at licensing and where the data came from.
Giving models your information
RAG Tools
RAG Tools help ground a model in your own data, so answers draw on your documents. Useful for internal assistants and support bots.
Vector Databases
Vector Databases store embeddings and run semantic search, the layer underneath most RAG setups.
Knowledge Base and Document Analysis
Knowledge Base covers company wikis and internal search. Document Analysis covers chatting with PDFs and extracting information from documents.
Building and orchestrating
AI Frameworks
AI Frameworks are libraries for building AI applications and agents. They speed up the first version, but adding one is a commitment, so check how active the project is.
AI Agents
AI Agents are autonomous systems that carry out tasks and make decisions. Decide what an agent may do without asking first, and keep a person in the loop for anything costly or irreversible.
MCP Servers
MCP Servers are servers and clients that connect AI assistants to tools and data via the Model Context Protocol. Only connect servers you trust, since they can read or change what they're given access to.
Prompt Engineering
Prompt Engineering tools write, test and manage prompts.
Automation
Automation covers workflow tools that connect apps and run tasks on triggers.
Running it
MLOps
MLOps tools train, deploy and monitor machine learning models. They matter once your project has real users and needs monitoring, versioning and cost control.
Robotics
Robotics lists tools and platforms for robots and embodied intelligence.
A rough path
- Pick a model provider (LLM Platforms).
- If answers must come from your data, add RAG Tools and a Vector Database.
- Write and test prompts (Prompt Engineering).
- Use a framework if the project needs one.
- Add monitoring before launch (MLOps).
Tool pages describe each product briefly. The details that decide a choice, like pricing, limits and privacy, are on the tool's own website.
