Tools

Claude models (Haiku, Sonnet, Opus) through the Messages API, with tool use, prompt caching and batches.

Official docs for Claude API (Anthropic)Lesson: Calling an LLM API for the first timeLesson: Prompt caching and token pricingLesson: Function and tool calling basics

GPT models through the Responses API, plus embeddings, batch and moderation endpoints.

Official docs for OpenAI APILesson: Calling an LLM API for the first timeLesson: Embeddings and cosine similarity

Google's Gemini models for text, images, audio and video.

Official docs for Gemini API (Google)Lesson: Calling an LLM API for the first timeLesson: Multimodal inputs

The main hub for open models, datasets and demo apps (Spaces), with hosted inference options.

Official docs for Hugging Face HubLesson: Open vs closed models, small vs largeLesson: Serverless functions and the easy hosting route

Run open models on your own laptop or server with one command and a local HTTP API.

Official docs for OllamaLesson: Ollama for local modelsLesson: Open vs closed models, small vs large

High-throughput serving engine for LLMs with PagedAttention, continuous batching and an OpenAI-compatible API.

Official docs for vLLMLesson: Serving models with vLLMLesson: GPU instances

One Python interface and proxy for 100+ model providers, with routing, fallbacks and cost tracking.

Official docs for LiteLLMLesson: Model routing, fallbacks and LLM gateways

The standard Python library for classic machine learning (regression, trees, clustering, metrics).

Official docs for scikit-learnLesson: Classic machine learning in one lessonLesson: Overfitting and train, validation and test splits

Fast arrays and maths in Python; the base of almost every AI library.

Official docs for NumPyLesson: Vectors and matrices, just enough for AILesson: Python for AI engineers

Tables (DataFrames) for loading, cleaning and analysing data in Python.

Official docs for pandasLesson: Python for AI engineers

OpenAI's fast tokenizer library, handy for counting tokens.

Official docs for tiktokenLesson: Tokenization

Parameter-efficient fine-tuning methods such as LoRA and QLoRA.

Official docs for PEFT (Hugging Face)Lesson: Fine-tuning with LoRA and QLoRA

Vector similarity search inside Postgres, with HNSW and IVFFlat indexes.

Official docs for pgvectorLesson: Vector databases, HNSW and pgvectorLesson: Access control in RAG

A dedicated vector database with filtering, hybrid search and a managed cloud.

Official docs for QdrantLesson: Vector databases, HNSW and pgvector

A fully managed vector database.

Official docs for PineconeLesson: Vector databases, HNSW and pgvector

A simple embedded vector store popular for prototypes.

Official docs for ChromaLesson: Vector databases, HNSW and pgvector

A small Python implementation of BM25 keyword ranking.

Official docs for rank-bm25Lesson: BM25 keyword search and hybrid searchLesson: Capstone: build a small RAG app from scratch

Parses PDFs, Office files, HTML and images into clean text and elements for ingestion.

Official docs for UnstructuredLesson: Document parsing and ingestion pipelines

An open protocol for connecting AI apps to tools and data through MCP servers.

Official docs for Model Context ProtocolLesson: MCP: servers and clientsLesson: MCP server trust

A library for building stateful agent workflows as graphs, with checkpoints and human-in-the-loop.

Official docs for LangGraphLesson: LangGraph, state and checkpointsLesson: Human in the loop

A lightweight Python framework for agents with tools, handoffs, guardrails and tracing.

Official docs for OpenAI Agents SDKLesson: The OpenAI Agents SDK and other frameworks

An open protocol for agents from different vendors to discover each other and exchange tasks.

Official docs for Agent2Agent (A2A) protocolLesson: The A2A protocol

Open-source LLM tracing, prompt management, evals and cost tracking.

Official docs for LangfuseLesson: Langfuse and observability dashboardsLesson: Tracing and OpenTelemetry for LLMsLesson: Prompt versioning, model versioning and registries

Metrics for evaluating RAG pipelines, such as faithfulness and context precision.

Official docs for RagasLesson: RAG evals: faithfulness, relevance and context precisionLesson: Eval tools at a glance

A pytest-style framework for LLM evals with many built-in metrics.

Official docs for DeepEvalLesson: Eval tools at a glanceLesson: Regression evals in CI and A/B testing prompts

Config-driven evals and red teaming for prompts and models, runnable in CI.

Official docs for promptfooLesson: Eval tools at a glanceLesson: Red teaming

The open standard for traces, metrics and logs, including GenAI semantic conventions.

Official docs for OpenTelemetryLesson: Tracing and OpenTelemetry for LLMsLesson: Logging and monitoring

Meta's open classifier models for unsafe content (Llama Guard) and prompt injection or jailbreaks (Prompt Guard).

Official docs for Llama Guard and Prompt GuardLesson: Input guardrails, output validation and moderationLesson: Prompt injection, indirect injection and jailbreaks

Detects and anonymises personal data (PII) in text and images.

Official docs for PresidioLesson: PII redaction and data leakage

Publishes the Top 10 for LLM applications and for agentic applications, plus guides.

Official docs for OWASP GenAI Security ProjectLesson: The OWASP Top 10 for LLM apps

An LLM vulnerability scanner that probes models for jailbreaks, leaks and other failures.

Official docs for garakLesson: Red teaming

Microsoft's toolkit for automated red teaming of generative AI systems.

Official docs for PyRITLesson: Red teaming

Scans git repositories and commits for leaked secrets such as API keys.

Official docs for gitleaksLesson: Secrets management and API key rotation

A Python web framework with type-hint validation and automatic OpenAPI docs.

Official docs for FastAPILesson: REST APIs with FastAPILesson: Capstone: build a small RAG app from scratch

Data validation with Python type hints, plus Pydantic Settings for configuration.

Official docs for PydanticLesson: Input validation with Pydantic, and configLesson: Structured outputs

The popular open-source relational database; with pgvector it also stores embeddings.

Official docs for PostgreSQLLesson: SQL vs NoSQL, indexing and transactions

An in-memory store for caching, rate limits, sessions and queues.

Official docs for RedisLesson: Caching with RedisLesson: Rate limiting

A widely used Python task queue for background jobs.

Official docs for CeleryLesson: Message queues and background jobs

Python's most popular testing framework.

Official docs for pytestLesson: Testing APIs with pytest

Object storage for files, with presigned URLs and event notifications.

Official docs for Amazon S3Lesson: File uploads to S3

Version control for code, prompts and configuration.

Official docs for GitLesson: Git and GitHub for deployment

CI/CD workflows that run on every push and pull request.

Official docs for GitHub ActionsLesson: GitHub ActionsLesson: CI/CD pipelinesLesson: Eval gates in CI

Packages an app and its dependencies into a container image that runs the same everywhere.

Official docs for DockerLesson: DockerLesson: Docker Compose

Runs and scales containers across many machines, with health checks and rolling updates.

Official docs for KubernetesLesson: Kubernetes basicsLesson: Autoscaling and load balancing

Infrastructure as code tools that create and update cloud resources from configuration files.

Official docs for OpenTofu / TerraformLesson: Infrastructure as code

Metrics collection, dashboards and alerting.

Official docs for Prometheus and GrafanaLesson: Alerts and dashboards