GitHub Copilot

GitHub Copilot & SiteToMarkdown

Give Copilot Agents your real docs, not their training cutoff

SiteToMarkdown converts entire docs sites—like Tailwind CSS, Stripe, Next.js, OpenAI, and your own APIs—into a single, clean Markdown file you add to your workspace. GitHub Copilot Agents then use that file as high‑quality, up‑to‑date context when generating and editing code.

  • Ground Copilot in the latest Tailwind, Stripe, and framework APIs—no more outdated suggestions.
  • Provide one Markdown file per library or service for easy #‑mention context.
  • Use context engineering best practices from VS Code: curated, concise, always-on documentation.

No subscriptions — pay only when you convert a site.

VS Code GitHub Copilot Chat adding Markdown docs as context

Copilot is only as good as the context you give it

GitHub Copilot Chat and Agents can pull in workspace context, use #‑mentions for files, and even #fetch URLs. But:

  • Docs live across dozens or hundreds of pages (Tailwind, Stripe, Next.js, OpenAI, etc.). Fetching one URL at a time doesn’t give full coverage.
  • Docs sites are often JS‑rendered, or block bots—Copilot’s web tools may not see full or current content.
  • Manual copying into your repo is tedious, noisy (HTML, nav, ads), and bloats Copilot’s context window.

SiteToMarkdown gives you a single, clean Markdown file per docs site so Copilot Agents always have a concise, comprehensive reference in your workspace.

Copilot Chat context features you can combine with SiteToMarkdown

  • #file-or-folder — reference your Markdown docs directly.
  • #codebase — let Copilot search implementation code that uses those libraries.
  • @workspace — ask workspace‑level questions grounded in your imported docs.

Based on VS Code’s “Manage context for AI” and “Context engineering” guides .

A simple flow: convert docs → add to repo → reference in Copilot chat

Use SiteToMarkdown as the “context loader” in your Copilot context engineering workflow. Curate the docs once, then let Copilot Agents reuse them across chats and tasks.

Step 1

Convert library docs with SiteToMarkdown

Paste the docs URL into SiteToMarkdown:

  • https://tailwindcss.com/docs/
  • https://stripe.com/docs/api
  • https://nextjs.org/docs
  • https://platform.openai.com/docs/overview

SiteToMarkdown crawls all pages, renders JS, bypasses basic blocking, and merges the main content into a single Markdown file—no HTML, nav, or ads.

Step 2

Add the Markdown file to your project

Commit the exported files into a docs/ or .github/ folder:

docs/
  tailwindcss.md
  stripe-api.md
  nextjs.md
.github/
  copilot-instructions.md

Reference these docs from .github/copilot-instructions.md so Copilot Agents see them as project‑wide context.

Step 3

Use Copilot Chat with curated docs context

In Copilot Chat, use #‑mentions and @workspace to bring your docs into the prompt:

"Suggest card layout classes for this component based on #docs/tailwindcss.md"

"Using #docs/stripe-api.md, implement a payment intent endpoint in this file"

"@workspace explain how authentication works in our API using #docs/openai-api.md"

Copilot’s context window focuses on clean Markdown, not noisy web pages, improving accuracy and reducing back‑and‑forth.

Use SiteToMarkdown in your Copilot context engineering workflow

VS Code’s context engineering guide recommends keeping curated project‑wide context in Markdown files and referencing them from .github/copilot-instructions.md. SiteToMarkdown makes it trivial to do this for external libraries and APIs you depend on.

Example: Tailwind CSS + Stripe in a Next.js project

Suppose your project uses Tailwind CSS for styling, Stripe for billing, and Next.js for routing:

  • Convert Tailwind docs → docs/tailwindcss.md
  • Convert Stripe API docs → docs/stripe-api.md
  • Convert Next.js docs → docs/nextjs.md

Then add them to your Copilot instructions:

# Project context

- Refer to [Tailwind CSS utilities](../docs/tailwindcss.md) for styling.
- Use [Stripe API](../docs/stripe-api.md) for payments; follow their latest best practices.
- Follow [Next.js routing conventions](../docs/nextjs.md) for pages and API routes.

If any generated code conflicts with these docs, prefer the docs.

Result: Copilot that “knows” your stack as it is today

With this setup, Copilot Agents can:

  • Generate Tailwind classes that match the latest utilities and responsive variants.
  • Use the current Stripe API endpoints, parameters, and webhooks flows.
  • Follow Next.js’ current routing, app directory, and server component patterns.
  • Search your codebase (#codebase) while grounded in the correct docs.

Over time, you can treat these Markdown files as living “external context layers” alongside your PRODUCT/ARCHITECTURE/CONTRIBUTING docs recommended in the context engineering guide.

Prompt examples for Copilot Chat with SiteToMarkdown docs

Once your Markdown docs are in the repo, you can explicitly bring them into Copilot Chat using #‑mentions and chat participants.

Tailwind CSS

"Using #docs/tailwindcss.md, propose responsive classes
for this layout so that cards don't break across columns."

"@workspace refactor this component's styling to use
Tailwind utilities based on our Tailwind docs in
#docs/tailwindcss.md."

Stripe API

"Based on #docs/stripe-api.md, implement an endpoint that
creates a PaymentIntent for this checkout flow."

"Review our existing Stripe integration code against
#docs/stripe-api.md and list any deprecated fields or
endpoints we are still using."

Next.js + OpenAI

"Using #docs/nextjs.md, migrate our pages router API route
to the app router convention."

"With #docs/openai-api.md as reference, add a streaming
chat endpoint that follows OpenAI's latest Responses API."

Custom APIs & internal docs

"Following our internal API spec in #docs/internal-api.md,
add a new endpoint and update the client SDK."

"@workspace check that the authentication flow in this
feature is consistent with #docs/auth-spec.md."