A practical library for building with LLM chat APIs.

Getting started

Build a reliable chat assistant

Define the user task, build a minimal message loop and add streaming, error handling and evaluation in stages.

Choose the right chat model

Compare providers with a repeatable dataset and objective success criteria.

RAG for AI chat applications

Retrieve useful evidence and make grounded answers easier to verify.

Manage LLM chat API costs

Understand token budgets, context growth, caching and cost per successful task.

Design safe chat agents

Control tool access, confirmations and failures in action-oriented assistants.

Documentation companion

Use the documentation index for request structure, streaming, tool calls and structured output patterns.