LLM-ready Docs

Working with AI tools requires high-quality context. DatoCMS addresses this by providing documentation in formats optimized for AI consumption, allowing assistants like Claude, ChatGPT, Cursor, and NotebookLM to deliver accurate, context-aware responses about DatoCMS features and APIs.

DatoCMS offers three complementary approaches to accessing documentation for AI tools:

Complete documentation (llms-full.txt)

The complete DatoCMS documentation compiled into a single, perfectly formatted Markdown file. This includes all 500+ pages of content: API references, guides, migrations, plugins, Content Management API (CMA), Content Delivery API (CDA), and more.

Access:

• URL: https://www.datocms.com/docs/llms-full.txt

What makes it effective:

• Clean Markdown with properly formatted code blocks

• Logical structure maintained across all pages

• Complete context spanning the entire documentation

• Automatically regenerated with every docs update

• No navigation menus, JavaScript, or extraneous content

Use cases:

• Building complex features that require understanding multiple DatoCMS concepts

• Content migration projects

• Team onboarding and training

• Creating custom AI assistants with comprehensive DatoCMS knowledge

Documentation index (llms.txt)

A structured index of DatoCMS documentation following the llms.txt standard. Provides an overview of available documentation without the full content.

Access:

• URL: https://www.datocms.com/docs/llms.txt

Use cases:

• Quick reference for documentation structure

• Discovering available topics and guides

• Navigation for AI tools that support llms.txt format

Single-page Markdown export

Every documentation and blog page includes a "Copy page" dropdown that provides content in AI-friendly formats. This enables quick extraction of specific pages without dealing with HTML formatting issues or web scraping complications.

Access methods:

• Copy as Markdown: Click the "Copy page" dropdown on any docs or blog page

• Direct URL conversion: Append .md to any page URL to retrieve its Markdown version

Example:

https://www.datocms.com/docs/content-management-api.md

Use cases:

• Troubleshooting specific integrations

• Exploring particular API methods

• Learning about individual DatoCMS features

• Providing focused context to AI assistants for targeted questions

Integrations

Claude Projects

Claude Projects allow you to attach custom knowledge to Claude conversations. This is particularly effective with llms-full.txt for comprehensive DatoCMS expertise.

Setup:

1. Navigate to claude.ai and create a new Project

2. Download llms-full.txt from https://www.datocms.com/docs/llms-full.txt

3. Upload the file to your Project

4. Add custom instructions (see reference instructions)

5. Name your Project (e.g., "DatoCMS Docs")

Capabilities:

• Ask migration questions and receive step-by-step instructions

• Request working TypeScript scripts for content operations

• Plan complex content model migrations with full DatoCMS context

• Get accurate answers across all conversations in the Project

Custom GPTs

Build a ChatGPT assistant specialized in DatoCMS using the complete documentation as its knowledge base.

Setup:

1. Go to ChatGPT GPT Builder

2. Click "Create a GPT"

3. Download llms-full.txt from https://www.datocms.com/docs/llms-full.txt

4. In the Knowledge section, upload the downloaded file

5. Add instructions such as: "You are a DatoCMS expert. Answer questions using only the provided documentation. Include code examples when relevant." (see reference instructions)

Reference implementation:

Check the official DatoCMS Expert GPT to see a working example.

NotebookLM

Google's NotebookLM excels at deep research and learning across large documentation sets.

Setup:

1. Create a new notebook in NotebookLM

2. Add a source → paste https://www.datocms.com/docs/llms-full.txt

3. Allow processing to complete (~30 seconds)

Capabilities:

• Compare different DatoCMS features and APIs

• Generate study guides for learning specific topics

• Search across the entire documentation for related concepts

• Understand relationships between different parts of the system

Use cases:

• Onboarding new team members

• Exploring unfamiliar features

• Research before implementing complex features

Cursor

Cursor is an AI-powered code editor that benefits from documentation context while coding.

Setup:

1. Open Cursor and type @Docs

2. Select "Add new doc"

3. Paste: https://www.datocms.com/docs/llms-full.txt

Important: Always type the @ symbol manually in the chat interface. Copy-pasting breaks the context reference.

Capabilities:

• Generate Next.js pages that fetch DatoCMS content

• Add pagination, filtering, or sorting to GraphQL queries

• Debug queries with full knowledge of available fields and filters

• Write TypeScript scripts that interact with the Content Management API

Windsurf

Windsurf is another AI coding assistant that supports custom documentation sources.

Setup:

1. Open settings → Documentation

2. Add new documentation source

3. Enter the URL: https://www.datocms.com/docs/llms-full.txt

Important: Always type the @ symbol manually in the chat interface. Copy-pasting breaks the context reference.

Capabilities:

• Same as Cursor: context-aware code generation for DatoCMS integrations

• API-aware debugging and query construction

Other AI tools

Most AI assistants that accept file uploads or URL references can use llms-full.txt:

• File upload tools: Download the file and upload directly

• URL-based tools: Reference https://www.datocms.com/docs/llms-full.txt

• Chat interfaces: Copy and paste relevant sections as needed

Best practices

Choosing the right format

Use single-page Markdown export when:

• You need information about a specific feature or API method

• Working on a focused task that doesn't require broader context

• You want to minimize token usage in your AI assistant

• Troubleshooting a specific error or implementation detail

Use llms-full.txt when:

• Building complex features that span multiple DatoCMS concepts

• Planning migrations or major content model changes

• Creating a persistent AI assistant with comprehensive DatoCMS knowledge

• Team members need to learn DatoCMS from scratch

• Working on projects where understanding the full system architecture matters

Effective prompting

Provide clear, specific prompts that take advantage of the documentation context:

Good examples:

• "Using the Content Management API, write a script to bulk-update all blog posts to add a new field"

• "How do I migrate content from WordPress to DatoCMS while preserving relationships between posts and categories?"

• "Create a Next.js component that fetches localized content and handles fallbacks according to DatoCMS best practices"

Less effective examples:

• "How do I update posts?" (too vague)

• "Write me a migration script" (missing context about source and requirements)

• "Make a component" (no details about what it should do)

Iterative refinement

AI assistants work best with iterative feedback:

1. Start with a clear initial request

2. Review the generated code or answer

3. Provide specific feedback on what needs adjustment

4. Request modifications: "Add error handling" or "Include image optimization"

Most tasks succeed on the first attempt, with occasional minor refinements needed through follow-up prompts.

Token consumption considerations

The impact of using complete documentation (llms-full.txt) varies significantly depending on how your AI tool handles knowledge bases.

Tools with intelligent retrieval (Claude Projects, Custom GPTs, NotebookLM):

These systems don't load all 500+ pages into every conversation. Instead, they index the documentation and query only relevant sections as needed, retrieving context on-demand based on your questions. Token consumption is not a concern - the system automatically manages what context to include.

Tools without intelligent retrieval (direct chat interfaces, some coding assistants):

If you paste the full documentation directly into a chat or use tools that load entire files into context, you may encounter higher token usage per request and potential context window limits. However, the comprehensive context dramatically reduces incorrect or incomplete responses.

Limitations

• Markdown conversion quality: Quality may vary across different pages when using .md URL conversion

• Documentation updates: While llms-full.txt regenerates automatically, there may be a brief delay after documentation changes

• AI model capabilities: Results depend on the underlying AI model's capabilities and training

• Context windows: Some AI tools have limits on how much documentation they can process at once