Spring AI 2 - Intermediate

Learn metadata filtering to tag documents with author, date, topic, and year, then refine semantic vector search with filter expressions using and/or operators.

Build an etl pipeline for rag apps by ingesting documents from urls and file uploads, chunking with token text splitter for context limits, summarization and keywords enrichment, and security redaction.

Develop a basic ETL in a Spring AI app that ingests HTML pages by URL, extracts and chunks content, enriches with summaries and keywords, and stores vectors in batches.

Batch ingest processes multiple URLs in parallel with Java's completable future, ensuring failures don't cancel others, and loads extracted, chunked, and enriched documents into Redis vector store.

Learn to add custom metadata and content formatting to document chunks, using a default content formatter to exclude metadata from embedding and enable richer inference context.

Implement a custom document transformer in spring ai 2 intermediate to redact sensitive information and metadata, then apply summary and keyword enrichers in sequence before inference.

Learn how multimodal vision APIs understand images and text to describe images, extract text, compare images, and support applications from product photo analysis to document scanning and medical imaging.

Learn to build an API that compares two images, asks an LLM to find differences and similarities, and perform in-memory processing with byte array resources to avoid disk I/O.

Build an API that uses Spring AI structured output to convert receipt images into type-safe Java objects, extracting vendor, date, items, total, and notes.

Learn to generate images with Spring AI 2 by using a rest client to call Hugging Face's image API with a text prompt and a json payload.

Explore speech-to-text and text-to-speech in Spring AI, implementing an API to enable transcription and synthesis, enabling hands-free operation, notifications that talk, and podcast-like audible content.

Build a speech to text API that uploads an audio file, uses a whisper transcription model, and returns the transcribed text.

Develop an API that uploads a multi-part audio file, transcribes it to text via a transcription model, queries an LLM, then converts the reply to speech and returns the audio.

Explore voice vision analysis by uploading image and audio to an llm, receiving a text response, and converting it back to speech, demonstrated via a multipart API test.

Design a sequential workflow with specialized agents, researcher, writer, and editor, using chain, routing, and parallelization patterns in Spring AI to boost output quality.

Explore evaluator and optimizer workflows with reviewer and improver agents. Learn to iteratively refine drafts with prompts, feedback, and a final edit step, up to max iterations.

Learn a routing workflow that assigns user queries to distinct handlers by category, technical, analytical, or creative, via a classifier and dedicated agents.