dash-improve-my-llms 2.0

Crawler / SEO companion for Dash apps, with a thin MCP bridge.

What this package is

A small set of HTTP routes and middleware that make a Dash application legible to the parts of the web that don't speak Dash:

Search-engine and AI crawlers (Googlebot, GPTBot, ClaudeBot, …)
Users pasting a URL into a chat window for context
MCP-aware clients on Dash 4.3+

It is intentionally small. Dash 4.3 already ships an MCP server that exposes layouts and component metadata live over JSON-RPC, so this package no longer tries to compete on that surface.

The three audiences

Audience	How they talk to your app	What 2.0 gives them
MCP clients	JSON-RPC over Streamable HTTP	`LLMS_DOC` registered as `dash.mcp` resource
Web crawlers	Plain HTTPS, no JavaScript	`/robots.txt`, `/sitemap.xml`, static HTML
Paste-into-chat users	One-shot HTTP fetch	`/llms.txt`, `/<page>/llms.txt` as markdown

What it serves

/llms.txt and /<page>/llms.txt — prose markdown, from LLMS_DOC
/robots.txt — bot-class access policies via RobotsConfig
/sitemap.xml — generated from dash.page_registry
Static HTML prerender — served to crawlers that hit a normal page URL

That is the entire HTTP surface. There are no /page.json, /architecture.txt, or /llms.toon endpoints in 2.0 — component-tree introspection lives in Dash 4.3 MCP.

The LLMS_DOC pattern

Every page module exports a LLMS_DOC string. That string is the literal body of /<page>/llms.txt:

# pages/home.py
LLMS_DOC = '''
# Home
...
'''

If a page has no LLMS_DOC, the package emits a single warning at add_llms_routes() listing the missing pages, and the endpoint returns a small stub so bots still get a 200.

Multi-backend support

add_llms_routes(app) detects whether app.server is Flask, FastAPI, or Quart and dispatches to the matching adapter. Pick one extra at install time:

pip install dash-improve-my-llms[flask]     # Dash 3.x and earlier
pip install dash-improve-my-llms[fastapi]   # Dash 4.1+
pip install dash-improve-my-llms[quart]     # Dash 4.1+ async

Quick start

from dash import Dash
from dash_improve_my_llms import add_llms_routes, RobotsConfig, mark_hidden

app = Dash(__name__, use_pages=True)
app._base_url = "https://myapp.com"
app._robots_config = RobotsConfig(block_ai_training=True, allow_ai_search=True)

mark_hidden("/admin")
add_llms_routes(app)

Pages in this demo

Each of the three audience cards above has a dedicated demo page:

/audiences/mcp-clients — directory of resources this app registers with dash.mcp
/audiences/web-crawlers — live /robots.txt, RobotsConfig knobs, per-page visibility
/audiences/llm-context — copy any page's LLMS_DOC to your clipboard

Plus a few supporting pages:

/analytics — a regular Dash page (Plotly + callbacks) with its own LLMS_DOC
/admin — hidden via mark_hidden, shows visitor analytics
/v200-features — walks through what changed from 1.x

Note for AI agents: This is the static, prerendered view of an interactive Dash application served because we detected a non-JS user agent. Full prose docs:

/llms.txt — LLM-friendly documentation
/sitemap.xml
/robots.txt