Helidon Assistant is a Retrieval-Augmented Generation (RAG) application built around Helidon's documentation. It provides a friendly AI chat interface to help answer Helidon-related questions.
Key Features:
- Smart AsciiDoc Processing: Special handling of AsciiDoc content during embedding creation, including:
- Clean conversion to plain text.
- Grouping by document sections.
- Preserving code snippets and tables as meaningful units.
- Metadata Preservation: Embeddings retain rich metadata such as:
- Document name.
- Section and position within the document.
- Stateless Backend: Conversations are summarized on the server and stored client-side for a lightweight, scalable experience.
First, clone the Helidon GitHub repository to a temporary location on your local machine.
Update the project's application.yaml file to point to the Helidon documentation source, typically located at {helidon_repo_root}/docs/src/main/asciidoc. You can also define inclusion and exclusion filters as needed:
app:
root: "//{helidon_repo_root}/docs/src/main/asciidoc"
# exclusions:
inclusions: "*.adoc"Use Maven to build the application:
mvn clean packageLaunch the assistant with:
java -jar target/helidon-assistant.jarOnce the application is running, open your browser and navigate to http://localhost:8080 to start chatting with the Helidon Assistant.
Use the chat interface to ask questions and get helpful insights based on Helidon's official documentation.