Installation and setup

LlamaIndex installs as a core package plus one integration per model or store. Left unset, it calls OpenAI; this course uses local and stand-in models instead.

llama-index-core and two integrations

llama-index-core is the framework. The other two are integrations, installed separately as LlamaIndex does for every model and store: local embeddings through Hugging Face, and keyword search with BM25. Allow about 1 GB of disk for PyTorch and two small models.

pip install "llama-index-core==0.14.24" "llama-index-embeddings-huggingface==0.8.0" "llama-index-retrievers-bm25==0.8.0"

Example

from importlib.metadata import version

print(version("llama-index-core"))

An answering model from a free providerOptional

If no model is set, LlamaIndex uses OpenAI for answers and embeddings, which needs a paid key. The course sets a local embedding model in lesson 3 and a stand-in answering model in lesson 11, so neither default is used. A real answering model is one integration package and one line in Settings.

python

from llama_index.core import Settings
from llama_index.llms.groq import Groq

Settings.llm = Groq(model="openai/gpt-oss-120b")

Provider	Key	Variable	Package and class
Groq	Free, console.groq.com/keys	`GROQ_API_KEY`	`llama-index-llms-groq`, `Groq(model="openai/gpt-oss-120b")`
Gemini	Free, aistudio.google.com/apikey	`GOOGLE_API_KEY`	`llama-index-llms-google-genai`, `GoogleGenAI(model="gemini-2.5-flash")`
OpenRouter	Free models end in `:free`, openrouter.ai/keys	`OPENROUTER_API_KEY`	`llama-index-llms-openrouter`, `OpenRouter(model="google/gemma-4-31b-it:free")`
OpenAI	Paid, platform.openai.com/api-keys	`OPENAI_API_KEY`	`llama-index-llms-openai`, `OpenAI(model="gpt-4o-mini")`

pip install "llama-index-llms-groq==0.6.0"

export GROQ_API_KEY=gsk_...

Try it yourself

Run pip list and count the packages whose names start with llama-index.

PreviousLlamaIndex overview

Little by little, you're building something great.