Skip to content

Add /llms.txt to Sequelize documentation for better LLM integration #784

Open
@roychri

Description

@roychri

Issue Creation Checklist

Feature Description

Describe the feature you'd like to see implemented

I'm proposing to add an /llms.txt file to the Sequelize documentation website to make it more accessible to AI assistants and Large Language Models (LLMs).

The /llms.txt file would:

  1. Be written in Markdown format at the root of the documentation site
  2. Provide a concise overview of Sequelize
  3. Link to LLM-friendly (markdown) versions of key documentation pages
  4. Follow a specific format that's easily parsable by both humans and machines

Here's a sample of what a Sequelize llms.txt file might look like:

# Sequelize

> Sequelize is a modern TypeScript and Node.js ORM for PostgreSQL, MySQL, MariaDB, SQLite, MS SQL Server, and more. It features solid transaction support, relations, eager and lazy loading, read replication and more.

Sequelize allows you to model your database and the relationships between data programmatically, with features like automatic SQL query generation, schema validation, and database synchronization.

## Core Documentation

- [Getting Started](https://sequelize.org/docs/v6/getting-started.md): Introduction to Sequelize and basic setup
- [Model Basics](https://sequelize.org/docs/v6/core-concepts/model-basics.md): Creating models and understanding data types
- [Model Instances](https://sequelize.org/docs/v6/core-concepts/model-instances.md): Working with model instances
- [Model Querying - Basics](https://sequelize.org/docs/v6/core-concepts/model-querying-basics.md): Basic data retrieval operations
- [Associations](https://sequelize.org/docs/v6/core-concepts/assocs.md): Setting up and using model relationships
- [Transactions](https://sequelize.org/docs/v6/other-topics/transactions.md): Using transactions for data integrity

## Advanced Topics

- [Hooks](https://sequelize.org/docs/v6/other-topics/hooks.md): Using lifecycle hooks
- [Migrations](https://sequelize.org/docs/v6/other-topics/migrations.md): Database migration strategies
- [Raw Queries](https://sequelize.org/docs/v6/core-concepts/raw-queries.md): Using raw SQL with Sequelize

## Optional

- [Resources](https://sequelize.org/docs/v6/other-topics/resources.md): Additional learning resources
- [TypeScript](https://sequelize.org/docs/v6/other-topics/typescript.md): TypeScript integration guide

Additionally, each documentation page would ideally have a markdown version available by appending .md to its URL.

Describe why you would like this feature to be added to Sequelize

As LLMs become increasingly integrated into development workflows, they need high-quality, structured information to provide accurate assistance. However, most LLMs struggle with:

  1. Context limitations: They can't process entire documentation sites
  2. HTML complexity: They often fail to correctly extract information from complex web pages
  3. Finding essential info: They need guidance on what's most important

The llms.txt standard addresses these issues by providing:

  • A concise project overview in a standardized format
  • Links to markdown versions of key documentation
  • A curated selection of the most important resources

This would significantly improve the experience for developers who use AI coding assistants when working with Sequelize, reducing incorrect information and improving the quality of assistance.

Several other projects have already implemented this standard, including the FastHTML project and various nbdev-based libraries.

The standard is documented here: https://llmstxt.org/ and you can see Anthropic's version here: https://docs.anthropic.com/llms.txt

A few curated directory of adoption are popping up like:

Is this feature dialect-specific?

  • No. This feature is relevant to Sequelize as a whole.
  • Yes. This feature only applies to the following dialect(s):

Would you be willing to resolve this issue by submitting a Pull Request?


Indicate your interest in the addition of this feature by adding the 👍 reaction. Comments such as "+1" will be removed.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions