IDE and AI Coding Tool Setup

Setup directives for using this repository's llms.txt files with every major AI coding tool and IDE. Consume the relevant section to configure context for the Bun/Elysia/htmx/Prisma/Tailwind/daisyUI/Better Auth stack.

All raw file URLs follow this pattern:

https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/<tool>/llms.txt

Full URL table:

Technology	Raw URL
Bun	`https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/bun/llms.txt`
ElysiaJS	`https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt`
Better Auth	`https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/better-auth/llms.txt`
htmx	`https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt`
Prisma	`https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/prisma/llms.txt`
Tailwind CSS	`https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/tailwindcss/llms.txt`
daisyUI	`https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/daisyui/llms.txt`

GitHub Copilot in VS Code

Using `#fetch` in Copilot Chat

Type #fetch followed by a raw URL in the Copilot Chat input. Copilot fetches the URL and includes the content as context.

Example prompt:

#fetch https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt
How do I use the htmx sse extension to receive server-sent events?

Fetch multiple files in one prompt:

#fetch https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt
#fetch https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt
Show me how to send an htmx partial response from an Elysia route.

Auto-including context with `copilot-instructions.md`

Create .github/copilot-instructions.md in your project root. Add a section listing the raw URLs. Copilot reads this file automatically on every chat session.

Example .github/copilot-instructions.md:

## External Documentation

When working on this project, reference these documentation files for the technology stack:

- Bun runtime and bundler: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/bun/llms.txt
- ElysiaJS web framework: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt
- Better Auth integration: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/better-auth/llms.txt
- htmx hypermedia library: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt
- Tailwind CSS: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/tailwindcss/llms.txt
- daisyUI component library: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/daisyui/llms.txt
- Prisma ORM: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/prisma/llms.txt

Always use the Bun runtime (not Node.js) and Elysia for HTTP handling.
Use htmx attributes for client-side interactivity instead of a JavaScript framework.

Example prompts

#fetch https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/prisma/llms.txt
Create a Prisma $extends client extension that adds a softDelete method to all models.

#fetch https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/bun/llms.txt
Write a Bun plugin that loads YAML files as JavaScript objects.

GitHub Copilot in JetBrains IDEs

Using Copilot Chat in JetBrains

Install the GitHub Copilot plugin from JetBrains Marketplace. Paste a raw URL directly into the Copilot Chat input; Copilot fetches the URL and uses the content as context.

Example:

Read this documentation: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt
Now create an Elysia route that validates JWT tokens using the @elysiajs/jwt plugin.

Limitations compared to VS Code

The #fetch shorthand is not available in JetBrains; paste the full URL as plain text.
The copilot-instructions.md file is honoured by Copilot in JetBrains from Copilot plugin version 1.5 and later.
File attachment via the chat UI is not supported; use URL references instead.

Use the copilot-instructions.md file (same as VS Code) for persistent context at .github/copilot-instructions.md.

Cursor

Using `.cursor/rules/` directory

The .cursor/rules/ directory is the recommended approach for persistent rules and context (.cursorrules is deprecated). Create a .mdc file (Markdown with YAML frontmatter).

Example .cursor/rules/llms-stack.mdc:

---
description: Documentation context for the Bun/Elysia/htmx/Prisma/Tailwind/daisyUI stack
globs:
  - "**/*.ts"
  - "**/*.tsx"
  - "**/*.html"
  - "**/*.css"
alwaysApply: false
---

## Stack Documentation

When generating or reviewing code for this project, fetch and read the relevant documentation files:

- Bun: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/bun/llms.txt
- ElysiaJS: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt
- Better Auth: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/better-auth/llms.txt
- htmx: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt
- Prisma: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/prisma/llms.txt
- Tailwind CSS: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/tailwindcss/llms.txt
- daisyUI: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/daisyui/llms.txt

## Conventions

- Use Bun (not Node.js) as the runtime.
- Use Elysia for all HTTP routing.
- Use htmx attributes for interactivity. Do not add a JavaScript framework.
- Use Prisma for database access.
- Use Tailwind CSS and daisyUI for styling.

Using `@docs` in Cursor Chat

Type @docs in Cursor Chat to open the documentation picker. Add a documentation source by URL if not already added.

Using `@web` with raw GitHub URLs

In Cursor Chat, type @web followed by a raw URL to fetch and include the file as context:

@web https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/bun/llms.txt
How do I register a Bun plugin for loading SVG files?

Windsurf (Codeium)

Configuring Windsurf rules

Add a rule in the Windsurf settings (Rules or Memory section) listing the raw documentation URLs:

When working on this project, reference the following documentation files for context:
- https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/bun/llms.txt
- https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt
- https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt
- https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/prisma/llms.txt
- https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/tailwindcss/llms.txt
- https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/daisyui/llms.txt
- https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/better-auth/llms.txt

Using Cascade with documentation context

In Cascade (Windsurf's agentic chat panel), paste a raw URL directly into the chat input. Cascade fetches the file.

Example:

Fetch this file and use it as context: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt
Create an Elysia application with JWT authentication and a protected route.

Attaching files in Windsurf Chat

Click the attachment icon in the Windsurf Chat input and select a locally cloned llms.txt file. The file contents are included as context for the current chat session.

Cline (VS Code extension)

Creating a `.clinerules` file

Create .clinerules at the project root with documentation instructions. Cline reads this file automatically when a task starts.

Example .clinerules:

## Documentation Context

When working on this project, fetch and read the following documentation files before answering:

- Bun: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/bun/llms.txt
- ElysiaJS: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt
- htmx: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt
- Prisma: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/prisma/llms.txt
- Tailwind CSS: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/tailwindcss/llms.txt
- daisyUI: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/daisyui/llms.txt
- Better Auth: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/better-auth/llms.txt

## Stack Conventions

- Runtime: Bun
- HTTP framework: Elysia
- Database ORM: Prisma
- UI interactivity: htmx (no JavaScript framework)
- Styling: Tailwind CSS + daisyUI
- Auth: Better Auth

Cline reads .clinerules automatically when a task starts in that workspace.

Using `@url` in Cline chat

In the Cline chat input, use @url to fetch a remote file on demand:

@url https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/prisma/llms.txt
Create a Prisma schema for a blog application with users and posts.

Configuring custom instructions

In VS Code settings, search for "Cline: Custom Instructions". Add the documentation URLs and stack conventions as persistent instructions.

Claude (via API or claude.ai)

Pasting raw URLs

Paste a raw URL directly into the Claude message input. Claude fetches the URL and includes the content as context.

Read this documentation file: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt

Now explain how to build a custom htmx extension that adds a loading spinner during requests.

Attaching downloaded files

Download the llms.txt file and attach it using the attachment button in claude.ai.

Using Claude Projects

Upload llms.txt files to a Project's knowledge base. All conversations in the Project will have access to the uploaded files.

Recommended files to upload for this stack:

bun/llms.txt
elysiajs/llms.txt
htmx/llms.txt
prisma/llms.txt
tailwindcss/llms.txt
daisyui/llms.txt
better-auth/llms.txt

System prompt pattern for API usage

You are a TypeScript full-stack developer using the Bun/Elysia/htmx/Prisma/Tailwind CSS/daisyUI stack.

The following documentation indexes are provided as context. Read each one before answering.

<bun_docs>
[contents of bun/llms.txt]
</bun_docs>

<elysia_docs>
[contents of elysiajs/llms.txt]
</elysia_docs>

[... continue for each tool ...]

ChatGPT / OpenAI

Pasting raw URLs

Paste a raw GitHub URL into the ChatGPT message input. ChatGPT will browse to the URL if browsing is enabled.

Read this file: https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt
Then create an Elysia server with a /users route that returns JSON.

Note: URL browsing availability depends on the ChatGPT model and configuration.

Using Custom GPTs with uploaded knowledge files

In the ChatGPT GPT editor, upload llms.txt files in the "Knowledge" section. Set instructions to reference the uploaded files when answering stack-related questions.

Example GPT instructions:

You are a full-stack TypeScript developer assistant specializing in the Bun/Elysia/htmx/Prisma/Tailwind/daisyUI stack.

When users ask about any of these technologies, refer to the uploaded documentation files for accurate API details.

Using Code Interpreter with attached files

Enable the Code Interpreter tool and attach a llms.txt file using the attachment button. ChatGPT can read the file content and use it in its response.

Aider

Using `--read` flag

Add a llms.txt file as read-only context. Aider includes the file in the LLM context but does not allow the LLM to edit it.

aider --read bun/llms.txt src/server.ts

Pass multiple files:

aider \
  --read bun/llms.txt \
  --read elysiajs/llms.txt \
  --read htmx/llms.txt \
  src/server.ts src/routes.ts

Configuring `.aider.conf.yml`

Create .aider.conf.yml at the project root to always include documentation files:

read:
  - /path/to/llms-stack-refresh/bun/llms.txt
  - /path/to/llms-stack-refresh/elysiajs/llms.txt
  - /path/to/llms-stack-refresh/htmx/llms.txt
  - /path/to/llms-stack-refresh/prisma/llms.txt
  - /path/to/llms-stack-refresh/tailwindcss/llms.txt
  - /path/to/llms-stack-refresh/daisyui/llms.txt
  - /path/to/llms-stack-refresh/better-auth/llms.txt

Replace /path/to/llms-stack-refresh/ with the actual path to your local clone of this repository.

Example aider commands

# Ask a question with documentation context
aider --read bun/llms.txt --message "How do I write a Bun plugin that loads .env files?"

# Edit a file with documentation context
aider --read elysiajs/llms.txt src/app.ts

Continue (VS Code / JetBrains extension)

Configuring `.continue/config.json`

Create or edit .continue/config.json in your project root. Add a contextProviders entry with a url provider listing the raw URLs.

Example .continue/config.json:

{
  "models": [],
  "contextProviders": [
    {
      "name": "url",
      "params": {
        "urls": [
          "https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/bun/llms.txt",
          "https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/elysiajs/llms.txt",
          "https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/htmx/llms.txt",
          "https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/prisma/llms.txt",
          "https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/tailwindcss/llms.txt",
          "https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/daisyui/llms.txt",
          "https://raw.githubusercontent.com/d4551/llms-stack-refresh/main/better-auth/llms.txt"
        ]
      }
    }
  ]
}

Using `@docs` context provider

In Continue Chat, type @docs to open the documentation picker and select a configured URL context provider.

Setting up custom slash commands

Add a custom slash command in .continue/config.json that references the stack documentation:

{
  "slashCommands": [
    {
      "name": "stack-docs",
      "description": "Include full stack documentation as context",
      "prompt": "I am working on a Bun/Elysia/htmx/Prisma/Tailwind/daisyUI project. The documentation for each tool is available at these URLs: [list URLs]. Please fetch and read the relevant documentation before answering."
    }
  ]
}

Zed AI

Using the `/file` command

Clone this repository locally. In the Zed AI assistant panel, type /file followed by the path to a local llms.txt file.

Example:

/file /path/to/llms-stack-refresh/htmx/llms.txt
Explain how to use the htmx response-targets extension.

Configuring Zed assistant context

In Zed settings, navigate to the assistant section. Add a default_context or similar setting pointing to local documentation files.

Note: Zed's assistant configuration evolves rapidly. Check https://zed.dev/docs/assistant for the latest approach.

Local Development (general)

Cloning the repository

git clone https://github.com/d4551/llms-stack-refresh.git
cd llms-stack-refresh

Files are in: bun/, daisyui/, better-auth/, elysiajs/, htmx/, prisma/, tailwindcss/

Symlinking files into a project

Create a symlink from a project's docs directory to a specific llms.txt file:

mkdir -p my-project/docs/llms
ln -s /path/to/llms-stack-refresh/bun/llms.txt my-project/docs/llms/bun.txt
ln -s /path/to/llms-stack-refresh/elysiajs/llms.txt my-project/docs/llms/elysiajs.txt
ln -s /path/to/llms-stack-refresh/htmx/llms.txt my-project/docs/llms/htmx.txt

Using with any local LLM (Ollama, LM Studio, etc.)

Pass file contents directly to a local model via the command line:

# Single file
cat bun/llms.txt | ollama run llama3 "Using this documentation, how do I write a Bun plugin?"

# Multiple files concatenated
cat elysiajs/llms.txt htmx/llms.txt | ollama run llama3 "How do I return an htmx partial from an Elysia route?"

Use the OpenAI-compatible API to include file contents in a system message:

curl http://localhost:11434/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d "{
    \"model\": \"llama3\",
    \"messages\": [
      {
        \"role\": \"system\",
        \"content\": \"$(cat elysiajs/llms.txt)\"
      },
      {
        \"role\": \"user\",
        \"content\": \"How do I add CORS to an Elysia app?\"
      }
    ]
  }"

Keeping files up to date

Pull the latest documentation updates:

cd llms-stack-refresh
git pull origin main

FilesExpand file tree

IDE-SETUP.md

Latest commit

History

IDE-SETUP.md

File metadata and controls

IDE and AI Coding Tool Setup

GitHub Copilot in VS Code

Using #fetch in Copilot Chat

Auto-including context with copilot-instructions.md

Example prompts

GitHub Copilot in JetBrains IDEs

Using Copilot Chat in JetBrains

Limitations compared to VS Code

Cursor

Using .cursor/rules/ directory

Using @docs in Cursor Chat

Using @web with raw GitHub URLs

Windsurf (Codeium)

Configuring Windsurf rules

Using Cascade with documentation context

Attaching files in Windsurf Chat

Cline (VS Code extension)

Creating a .clinerules file

Using @url in Cline chat

Configuring custom instructions

Claude (via API or claude.ai)

Pasting raw URLs

Attaching downloaded files

Using Claude Projects

System prompt pattern for API usage

ChatGPT / OpenAI

Pasting raw URLs

Using Custom GPTs with uploaded knowledge files

Using Code Interpreter with attached files

Aider

Using --read flag

Configuring .aider.conf.yml

Example aider commands

Continue (VS Code / JetBrains extension)

Configuring .continue/config.json

Using @docs context provider

Setting up custom slash commands

Zed AI

Using the /file command

Configuring Zed assistant context

Local Development (general)

Cloning the repository

Symlinking files into a project

Using with any local LLM (Ollama, LM Studio, etc.)

Keeping files up to date

Using `#fetch` in Copilot Chat

Auto-including context with `copilot-instructions.md`

Using `.cursor/rules/` directory

Using `@docs` in Cursor Chat

Using `@web` with raw GitHub URLs

Creating a `.clinerules` file

Using `@url` in Cline chat

Using `--read` flag

Configuring `.aider.conf.yml`

Configuring `.continue/config.json`

Using `@docs` context provider

Using the `/file` command