dash-improve-my-llms

Make your Dash applications AI-friendly with automatic documentation, bot management, and SEO optimization.

---

🎯 Overview

dash-improve-my-llms is a comprehensive plugin that automatically generates five types of AI-friendly documentation and SEO resources for your Dash application:

Automatic Documentation (v0.1.0)

1. llms.txt - Comprehensive, context-rich markdown optimized for LLM understanding
2. page.json - Detailed technical architecture with interactivity and data flow
3. architecture.txt - ASCII art representation of entire application

Bot Management & SEO (v0.2.0 - NEW!)

4. robots.txt - Intelligent bot control with AI training bot blocking
5. sitemap.xml - SEO-optimized sitemap with intelligent priority inference
6. Static HTML - Bot-friendly pages with structured data

Privacy Controls (v0.2.0 - NEW!)

• mark_hidden() - Hide sensitive pages from AI bots and search engines
• Bot Detection - Differentiate between AI training, AI search, and traditional bots
• Configurable Policies - Fine-grained control over which bots can access what

---

🚀 Quick Start

Installation

pip install dash-improve-my-llms

Basic Setup (30 seconds)

from dash import Dash
from dash_improve_my_llms import add_llms_routes

app = Dash(__name__, use_pages=True)
add_llms_routes(app)  # That's it! 🎉

if __name__ == '__main__':
    app.run(debug=True)

Now visit:

• http://localhost:8050/llms.txt - LLM-friendly page context
• http://localhost:8050/page.json - Technical architecture
• http://localhost:8050/architecture.txt - App overview
• http://localhost:8050/robots.txt - Bot access control NEW!
• http://localhost:8050/sitemap.xml - SEO sitemap NEW!

---

✨ Key Features

📄 Automatic Documentation

• Three comprehensive formats (llms.txt, page.json, architecture.txt)
• Smart context extraction - Understands your app structure
• Callback tracking - Documents all data flows
• Component categorization - Automatic classification by purpose
• Navigation mapping - Tracks all internal/external links

🤖 Bot Management (NEW in v0.2.0)

• AI Training Bot Blocking - Block GPTBot, Claude-Web, CCBot, etc.
• AI Search Allowance - Allow ChatGPT-User, ClaudeBot, PerplexityBot
• Traditional Search Engines - Full support for Google, Bing, etc.
• Configurable Policies - Fine-grained control via RobotsConfig
• Bot Detection - Accurately identify bot types from user agents

🔒 Privacy Controls (NEW in v0.2.0)

• Hide Sensitive Pages - mark_hidden() excludes pages from AI/bots
• Component Hiding - Hide specific components from extraction
• Automatic Exclusion - Hidden pages removed from sitemaps/robots.txt
• 404 for Hidden Routes - Bots get 404 on hidden page docs

🌐 SEO Optimization (NEW in v0.2.0)

• Smart Sitemap Generation - Automatic priority inference
• Priority System - Homepage=1.0, Dashboards=0.9, Reports=0.8, Docs=0.7
• Change Frequency - Intelligent frequency detection (daily, weekly, monthly)
• Static HTML for Bots - Schema.org structured data, Open Graph tags
• Noscript Fallbacks - Content for non-JS crawlers

🧪 Fully Tested

• 88 comprehensive tests - 100% pass rate
• 98-100% coverage - All new modules fully tested
• Integration tests - Real-world scenario coverage
• Fast execution - 0.22s for entire test suite

---

📖 Complete Example

Setup with Bot Control

from dash import Dash, html, dcc
from dash_improve_my_llms import (
    add_llms_routes,
    mark_important,
    mark_hidden,
    register_page_metadata,
    RobotsConfig
)

# Create app
app = Dash(__name__, use_pages=True)

# Configure bot policies
robots_config = RobotsConfig(
    block_ai_training=True,      # Block GPTBot, CCBot, etc.
    allow_ai_search=True,         # Allow ClaudeBot, ChatGPT-User
    allow_traditional=True,       # Allow Googlebot, Bingbot
    crawl_delay=10,               # 10 second delay between requests
    disallowed_paths=["/admin", "/api/*"]  # Block specific paths
)

# Set base URL for SEO
app._base_url = "https://myapp.com"
app._robots_config = robots_config

# Add LLMS routes with all features
add_llms_routes(app)

# Hide admin pages from AI bots
mark_hidden("/admin")
mark_hidden("/settings")

# Add custom metadata for better SEO
register_page_metadata(
    path="/",
    name="Equipment Management System",
    description="Comprehensive equipment tracking and analytics platform"
)

if __name__ == '__main__':
    app.run(debug=True)

Page with Important Sections

# pages/equipment.py
from dash import html, Input, Output, callback
from dash_improve_my_llms import mark_important, register_page_metadata
import dash_mantine_components as dmc

register_page_metadata(
    path="/equipment",
    name="Equipment Catalog",
    description="Browse and filter the complete equipment catalog"
)

def layout():
    return html.Div([
        html.H1("Equipment Catalog"),

        # Mark filters as important for AI understanding
        mark_important(
            html.Div([
                html.H2("Filters"),
                dmc.TextInput(
                    id="equipment-search",
                    placeholder="Search equipment...",
                ),
                dmc.Select(
                    id="equipment-category",
                    data=[
                        {"value": "all", "label": "All Categories"},
                        {"value": "tools", "label": "Tools"},
                        {"value": "machinery", "label": "Machinery"},
                    ],
                    value="all"
                ),
            ], id="filters")
        ),

        html.Div(id="equipment-list"),
    ])

@callback(
    Output("equipment-list", "children"),
    Input("equipment-search", "value"),
    Input("equipment-category", "value"),
)
def update_list(search, category):
    # Your filtering logic here
    return html.Div("Equipment items...")

Hidden Admin Page

# pages/admin.py
from dash import html, register_page
from dash_improve_my_llms import mark_hidden

register_page(__name__, path="/admin", name="Admin Panel")

# This page won't appear in sitemaps or llms.txt
mark_hidden("/admin")

def layout():
    return html.Div([
        html.H1("Admin Panel"),
        html.P("Sensitive administrative controls")
    ])

---

🤖 Bot Management

RobotsConfig Options

from dash_improve_my_llms import RobotsConfig

# Default configuration (recommended)
config = RobotsConfig(
    block_ai_training=True,      # Block AI training bots
    allow_ai_search=True,         # Allow AI search bots
    allow_traditional=True,       # Allow traditional search engines
    crawl_delay=None,             # No delay
    custom_rules=[],              # No custom rules
    disallowed_paths=[]           # No additional blocks
)

# Strict configuration (block everything except Google)
strict_config = RobotsConfig(
    block_ai_training=True,
    allow_ai_search=False,
    allow_traditional=True,
    crawl_delay=30,
    disallowed_paths=["/admin", "/api", "/internal/*"]
)

# Open configuration (allow everything)
open_config = RobotsConfig(
    block_ai_training=False,
    allow_ai_search=True,
    allow_traditional=True
)

# Apply to app
app._robots_config = config

Bot Detection

The plugin automatically detects and handles different bot types:

Bot Type	Examples	Default Policy
AI Training	GPTBot, Claude-Web, CCBot, Google-Extended, anthropic-ai	❌ Blocked
AI Search	ChatGPT-User, ClaudeBot, PerplexityBot	✅ Allowed
Traditional	Googlebot, Bingbot, Yahoo, DuckDuckBot	✅ Allowed

from dash_improve_my_llms.bot_detection import (
    is_ai_training_bot,
    is_ai_search_bot,
    is_traditional_bot,
    get_bot_type
)

user_agent = "Mozilla/5.0 (compatible; GPTBot/1.0)"

if is_ai_training_bot(user_agent):
    print("AI training bot detected - blocking")

bot_type = get_bot_type(user_agent)  # Returns: "training", "search", "traditional", or "unknown"

---

🗺️ SEO Optimization

Sitemap Generation

The plugin automatically generates an SEO-optimized sitemap with intelligent priority inference:

# Automatic priority based on page type:
# - Homepage (/)           → Priority 1.0
# - Dashboards             → Priority 0.9
# - Reports/Analytics      → Priority 0.8
# - Documentation/Help     → Priority 0.7
# - Other pages            → Priority 0.5

# Change frequency inference:
# - Dashboards/Live        → daily
# - Reports/Analytics      → weekly
# - Documentation          → monthly
# - Static pages           → yearly

Example sitemap entry:

<url>
  <loc>https://myapp.com/</loc>
  <lastmod>2025-11-04</lastmod>
  <changefreq>weekly</changefreq>
  <priority>1.0</priority>
</url>

Bot Response Middleware (The Key Feature!)

Problem: AI crawlers cannot execute JavaScript, so they see empty <div id="react-entry-point"> placeholders instead of your actual content.

Solution: The middleware automatically detects bots and serves them llms.txt content wrapped in readable HTML.

# What bots receive:
✅ Search Bots (ClaudeBot, ChatGPT-User) → llms.txt content in HTML
✅ Traditional Bots (Googlebot, Bingbot)  → llms.txt content in HTML
❌ Training Bots (GPTBot, anthropic-ai)   → 403 Forbidden
✅ Regular Users (Chrome, Firefox)        → Full Dash React app

Before Middleware (❌ Bad):

<!-- Bots saw this - empty until JavaScript executes -->
<div id="react-entry-point">
    <div class="_dash-loading">Loading...</div>
</div>

After Middleware (✅ Good):

<!-- Bots now see this - readable content immediately -->
<div class="bot-notice">
    🤖 Bot-Optimized Content
    Also available: llms.txt | page.json | architecture.txt
</div>
<pre>
# Equipment Catalog

> Browse and filter the complete equipment catalog

## Key Content
- Equipment search and filtering
- Category selection
...
</pre>

Features:

• Automatic Detection: Identifies bot type from user agent
• Smart Serving: llms.txt content for bots, React app for users
• SEO Optimized: Includes Schema.org, Open Graph, meta tags
• Privacy Enforced: Training bots get 403 when blocked
• No JavaScript Required: Bots see content immediately

Static HTML Components

The HTML served to bots includes:

• Schema.org JSON-LD - Structured data for search engines
• Open Graph tags - Social media previews
• Meta tags - Description, robots, viewport
• Navigation links - Accessible site structure
• Bot notice banner - Links to documentation formats
• llms.txt content - Full page context in <pre> tag

---

🔒 Privacy Controls

Hiding Pages

from dash_improve_my_llms import mark_hidden, is_hidden

# Hide sensitive pages
mark_hidden("/admin")
mark_hidden("/settings")
mark_hidden("/internal/metrics")

# Check if page is hidden
if is_hidden("/admin"):
    print("Admin page is hidden from bots")

# Hidden pages are automatically:
# - Excluded from sitemap.xml
# - Blocked in robots.txt
# - Return 404 for /page-path/llms.txt
# - Return 404 for /page-path/page.json

Hiding Components

from dash_improve_my_llms import mark_component_hidden, is_component_hidden
from dash import html

# Hide sensitive components from extraction
api_key_display = html.Div([
    html.P("API Key: sk-..."),
    html.P("Secret: abc123"),
], id="api-keys")

mark_component_hidden(api_key_display)

# Check if component is hidden
if is_component_hidden("api-keys"):
    print("Component excluded from llms.txt")

---

📊 Generated Documentation

llms.txt (Comprehensive Context)

# Equipment Catalog

> Browse and filter the complete equipment catalog

## Application Context
This page is part of a multi-page Dash application with 3 total pages.

## Page Purpose
- **Data Input**: Contains form elements
- **Interactive**: Responds to user interactions

## Interactive Elements
**User Inputs:**
- TextInput (ID: equipment-search)
- Select (ID: equipment-category)

## Data Flow & Callbacks
**Callback 1:**
- Updates: equipment-list.children
- Triggered by: equipment-search.value, equipment-category.value

page.json (Technical Architecture)

{
  "path": "/equipment",
  "components": {
    "ids": {
      "equipment-search": {
        "type": "TextInput",
        "module": "dash_mantine_components"
      }
    },
    "categories": {
      "inputs": ["equipment-search", "equipment-category"],
      "interactive": ["equipment-search", "equipment-category"]
    }
  },
  "callbacks": {
    "list": [
      {
        "output": "equipment-list.children",
        "inputs": ["equipment-search.value"]
      }
    ]
  }
}

robots.txt (Bot Control)

# Robots.txt for Dash Application
# Block AI training bots, allow search bots

User-agent: GPTBot
Disallow: /

User-agent: anthropic-ai
Disallow: /

User-agent: ClaudeBot
Allow: /

User-agent: *
Allow: /
Crawl-delay: 10
Disallow: /admin
Disallow: /api/*

Sitemap: https://myapp.com/sitemap.xml

---

🧪 Testing

The package has comprehensive test coverage:

# Run all 88 tests
pytest tests/ -v

# Run with coverage
pytest tests/ --cov=dash_improve_my_llms --cov-report=term-missing

# Test results:
# ✅ Bot Detection: 14/14 tests (100% coverage)
# ✅ HTML Generator: 20/20 tests (100% coverage)
# ✅ Robots Generator: 16/16 tests (100% coverage)
# ✅ Sitemap Generator: 33/33 tests (98% coverage)
# ✅ Integration: 15/15 tests (Complete workflows)
# ✅ Total: 88/88 tests passing in 0.22s

See TEST_REPORT.md for detailed test documentation.

---

🎨 API Reference

Core Functions

add_llms_routes(app, config=None)

Add all LLMS routes to your Dash app (llms.txt, page.json, architecture.txt, robots.txt, sitemap.xml).

from dash_improve_my_llms import add_llms_routes, LLMSConfig

config = LLMSConfig(
    enabled=True,
    max_depth=20,
    include_css=True,
    include_callbacks=True
)

add_llms_routes(app, config)

mark_important(component, component_id=None)

Mark a component as important for LLM context. All children inherit importance.

important_section = mark_important(
    html.Div([...], id="key-metrics")
)

mark_hidden(page_path)

Hide a page from AI bots, search engines, and sitemaps.

mark_hidden("/admin")
mark_hidden("/settings")

register_page_metadata(path, name=None, description=None, **kwargs)

Register custom metadata for better SEO and documentation.

register_page_metadata(
    path="/analytics",
    name="Analytics Dashboard",
    description="Real-time business analytics",
    category="reporting"
)

Bot Management

RobotsConfig

Configuration for robots.txt generation.

Parameters:

• block_ai_training (bool): Block AI training bots (default: True)
• allow_ai_search (bool): Allow AI search bots (default: True)
• allow_traditional (bool): Allow traditional search engines (default: True)
• crawl_delay (int, optional): Delay between requests in seconds
• custom_rules (list, optional): Additional robots.txt rules
• disallowed_paths (list, optional): Paths to block

from dash_improve_my_llms import RobotsConfig

config = RobotsConfig(
    block_ai_training=True,
    crawl_delay=15,
    disallowed_paths=["/admin", "/api/*"]
)
app._robots_config = config

Bot Detection Functions

from dash_improve_my_llms.bot_detection import (
    is_ai_training_bot,
    is_ai_search_bot,
    is_traditional_bot,
    is_any_bot,
    get_bot_type
)

user_agent = request.headers.get('User-Agent', '')

# Check bot type
is_ai_training_bot(user_agent)  # Returns bool
is_ai_search_bot(user_agent)     # Returns bool
is_traditional_bot(user_agent)   # Returns bool
is_any_bot(user_agent)           # Returns bool
get_bot_type(user_agent)         # Returns "training", "search", "traditional", or "unknown"

---

🔧 Advanced Usage

Custom Sitemap Entries

from dash_improve_my_llms.sitemap_generator import SitemapEntry

custom_entry = SitemapEntry(
    loc="https://myapp.com/special",
    changefreq="monthly",
    priority=0.6
)

# Add to sitemap via configuration

Programmatic Access

from dash_improve_my_llms import (
    generate_llms_txt,
    generate_page_json,
    generate_architecture_txt
)
from dash_improve_my_llms.robots_generator import generate_robots_txt
from dash_improve_my_llms.sitemap_generator import generate_sitemap_xml

# Generate documentation programmatically
llms_content = generate_llms_txt("/mypage", layout_func, "My Page", app)
page_arch = generate_page_json("/mypage", layout_func, app)
app_arch = generate_architecture_txt(app)

# Generate SEO files
robots_content = generate_robots_txt(robots_config, sitemap_url, base_url)
sitemap_content = generate_sitemap_xml(pages, base_url)

---

🚀 Migration Guide

Upgrading from v0.1.0 to v0.2.0

v0.2.0 is fully backward compatible. All v0.1.0 code works without changes.

New features (optional):

# 1. Configure bot policies
app._robots_config = RobotsConfig(block_ai_training=True)

# 2. Set base URL for SEO
app._base_url = "https://myapp.com"

# 3. Hide sensitive pages
from dash_improve_my_llms import mark_hidden
mark_hidden("/admin")

# That's it! Enjoy:
# - /robots.txt
# - /sitemap.xml
# - Better SEO
# - Bot control

---

📦 What's New in v0.2.0

New Features

• ✅ Bot Detection - Identify AI training, AI search, and traditional bots
• ✅ Robots.txt Generation - Automatic with configurable policies
• ✅ Sitemap.xml Generation - Smart priorities and change frequencies
• ✅ Static HTML for Bots - Schema.org structured data
• ✅ Privacy Controls - mark_hidden() for sensitive pages
• ✅ Component Hiding - Exclude components from extraction

Improvements

• ✅ 88 Comprehensive Tests - 100% pass rate in 0.22s
• ✅ 98-100% Coverage - All new modules fully tested
• ✅ Better SEO - Priority inference, change frequency detection
• ✅ Bot Differentiation - Fine-grained control per bot type

Files Added

• dash_improve_my_llms/bot_detection.py - Bot user agent detection
• dash_improve_my_llms/robots_generator.py - robots.txt generation
• dash_improve_my_llms/sitemap_generator.py - sitemap.xml generation
• dash_improve_my_llms/html_generator.py - Static HTML for bots
• tests/test_bot_detection.py - 14 comprehensive tests
• tests/test_robots_generator.py - 16 comprehensive tests
• tests/test_sitemap_generator.py - 33 comprehensive tests
• tests/test_html_generator.py - 20 comprehensive tests
• tests/test_integration.py - 15 integration tests
• TEST_REPORT.md - Complete test documentation

---

📊 Compatibility

• Python: 3.8, 3.9, 3.10, 3.11, 3.12+
• Dash: 3.2.0+
• Dash Mantine Components: 2.3.0+ (optional)

Works with:

• ✅ Dash Pages (dash.register_page)
• ✅ Manual routing (dcc.Location)
• ✅ Multi-page apps
• ✅ Single-page apps
• ✅ All Dash component libraries

---

🤝 Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Add tests for new functionality
4. Ensure all 88 tests pass
5. Submit a pull request

# Run tests
pytest tests/ -v

# Run tests with coverage
pytest tests/ --cov=dash_improve_my_llms --cov-report=html

# Format code
black dash_improve_my_llms/ tests/

---

📄 License

MIT License - see LICENSE file for details.

---

🙏 Credits

Built by Pip Install Python LLC for the Dash community.

Inspired by:

• llms.txt specification
• dmc-docs

Special thanks to the Dash community and Plotly team.

---

🔗 Links

• Documentation: CLAUDE.md
• Test Report: TEST_REPORT.md
• PyPI: dash-improve-my-llms
• Dash: dash.plotly.com
• Plotly Pro: plotly.pro
• Issues: GitHub Issues

---

Made with ❤️ for the Dash community

⭐ Star on GitHub | 📖 Read the Docs | 🐛 Report Bug