Using AI with the Thredd Documentation Portal

Overview

This page explains how developers, tools, and AI systems can access and interpret the Thredd documentation efficiently. Whether you're using chatbots, code assistants, or building your own AI workflows, Thredd provides structured ways to make integration more efficient and reliable.

 

Using AI Tools with the Thredd Documentation

The Thredd Documentation Portal is designed to work well with modern AI tools such as:

  • Chat-based assistants (e.g., coding copilots)

  • Retrieval-Augmented Generation (RAG) systems

  • Internal knowledge bots

You can:

  • Paste documentation into AI tools

  • Point AI systems to the Thredd resources

  • Build pipelines that fetch and index the Thredd Documentation Portal automatically

Structured Content

The Thredd documentation is built with consistency to improve AI understanding:

  • Clear headings and hierarchy

  • Predictable page structure

  • Code examples separated from explanations

  • Minimal ambiguity in terminology

This improves:

  • Answer accuracy in AI tools

  • Chunking for embeddings

  • Semantic search performance

  • Programmatic Access (Optional Section)

Usage Guidelines

When using the Thredd documentation with AI systems:

  • Prefer official endpoints (llm-v1.txt, sitemap)

  • Keep content up to date (re-sync regularly)

  • Validate critical outputs against source docs

  • Avoid relying on outdated cached content

  • Do not assume undocumented behaviour

Best Practices

  • Use smaller chunks (300–800 tokens) for better retrieval

  • Keep metadata (For example, URLs and titles) attached to each chunk

  • Re-index periodically to reflect updates

  • Combine with your internal context if needed

Sitemap Access

Thredd provides a range of sitemaps to help both humans and machines discover all available content.

Why it matters:

  • Lists all documentation pages in a structured format

  • Helps AI systems crawl and index content efficiently

  • Useful for building automated ingestion pipelines

Sitemap Master Index

Maps all sitemaps across the docs.thredd.com domain.

Location:

https://docs.thredd.com/sitemapindex/Sitemap.xml

Merged Master Sitemap

Merges all of the docs sitemaps across the docs.thredd.com domain into a single sitemap.

Location:

https://docs.thredd.com/sitemap-all.xml

LLM-Friendly Content (llm-v1.txt)

Thredd provides machine-friendly indexes of the Thredd Documentation Portal for AI systems. Thredd provides both a full coverage LLM file and a high-priority LLM file.

What they include:

  • Curated list of important documentation pages

  • Clean, minimal structure optimised for LLM ingestion

  • Reduced noise compared to full HTML pages

Recommended usage:

  • Use as an entry point for AI ingestion

  • Feed into embedding pipelines

  • Use for RAG systems instead of crawling the full site

LLM File

This is a full map of the Thredd documentation, offering:

  • Broad coverage

  • Includes all major sections

  • Some deep links

This is ideal for:

  • General AI browsing

  • Search engines

  • Unknown queries

This file is version controlled and the current version is llm-v1.txt.

Location:

https://docs.thredd.com/llm-v1.txt

LLM High Priority File

Purpose:

  • Fast lane for answering real questions

  • Only the most useful, high-signal content

This file focusses on:

  • high traffic pages

  • API usage

  • common tasks

  • Minimal noise

This is optimal for:

  • RAG pipelines

  • chat assistants

  • internal AI tools

Location:

https://docs.thredd.com/llm-high-priority.txt

Example Usage

Using the Thredd Documentation in a Retrieval-Augmented Generation (RAG) Pipeline

  1. Fetch llm-v1.txt

  2. Retrieve linked pages

  3. Chunk content into sections

  4. Generate embeddings

  5. Store in a vector database

  6. Query using semantic search

Feedback

If you're building AI integrations for the Thredd documentation, Thredd would be interested to hear from you. Your feedback helps us improve machine readability and developer experience. If you want to contact the Thredd Technical Publications team directly, you can email us at: docs@thredd.com.