ANDHUB / AI RESOURCES

Give your tools
the right context.

Use the current company details, proposed brand standards and existing assets when asking an AI tool to write, design or build for AndHub. Choose the format your tool supports.

Download brand context JSON Download plain text TXT

01 / Read or attach

Start with the files.

Supply the discovery URL, attach the text summary, or read the structured JSON. No integration setup is needed for tools that support those inputs.

Choose a source →

02 / Reuse instructions

Use the brand skill.

A small reusable skill helps assistants choose the current source and appropriate assets. Keep it with the project or import it using your tool’s supported workflow.

Read the skill →

03 / Local MCP

Connect the library.

The read-only adapter exposes current context and catalog search to tools that support a locally launched stdio MCP server.

See local setup →

Source status matters. Company details come from the supplied current data. The brand standards remain proposed, and original logo artwork remains reconstructed for review. Historical PDF/DOCX wording predates the current vector and font library. Current user instructions take precedence.

Readable, structured, discoverable.

Start at https://brand.1hubengineering.com/llms.txt. If your tool cannot retrieve URLs, download the JSON or text and attach it to your prompt.

Resource	Use
llms.txt	Short discovery index with canonical source links and status.
llms-full.txt	Current company copy, contacts, brand rules and source priority as plain text.
brand-context.json	Typed context for automation. Its JSON Schema describes the data.
Complete guideline Markdown	All source-guide text and tables with a current-status preface and illustration links.
brand.json	Current supplied company facts, contacts, leadership and configured social destinations.
catalog.json	Resource IDs, categories, files, formats, previews, sizes and checksums. Relative paths resolve against the published project URL.
Organization JSON-LD	Schema.org company, address and founder data. Social URLs are configured destinations, not verified-account claims.

Example request

Use the attached AndHub brand context and the current asset catalog to prepare a company introduction. Use the supplied contact details and existing logo artwork. Keep the tone precise and approachable. Do not invent customers, certifications or results.

Automatic discovery of llms.txt varies by tool. Providing a direct link or attachment is the most explicit way to supply the material.

Reusable AndHub brand skill.

The andhub-brand skill contains scoped instructions and links to current sources. It helps with AndHub copy, interfaces, templates and resource selection.

Download the skill’s SKILL.md or take the assets/ai/andhub-brand/ directory from the repository or AI bundle.
Keep the folder with the project, or import it using the skill workflow supported by your assistant. No global configuration is changed by this website.
Ask your tool to use that skill for the AndHub task. In a local checkout, current facts are available in assets/brand.json and assets/ai/brand-context.json.

The root AGENTS.md supplies repository guidance for coding assistants that recognize it. The skill and repository instructions preserve the user’s current scope and permissions.

A real read-only local MCP adapter.

The adapter implements newline-delimited JSON-RPC over stdio for MCP 2025-11-25. It reads fixed local brand sources and the catalog, returns published download URLs, and offers no write tools or arbitrary file access.

Download a complete repository checkout or the portable AI bundle. Keep its folder structure so assets/ai/mcp-server.mjs can find the brand context and resource index.
Have Node.js 22 or newer available. The adapter uses the Node standard library and needs no package installation.
Add the following entry through your MCP client’s supported local-server settings. Replace the placeholder with the actual absolute file path.

{
  "mcpServers": {
    "andhub-brand": {
      "command": "node",
      "args": ["/PATH/TO/branding-guideline/assets/ai/mcp-server.mjs"]
    }
  }
}

On Windows, use forward slashes or escaped backslashes in the JSON path. If your client cannot find node, use the full path to its Node executable as the command. Configuration shapes differ between clients; the example shows a common local stdio layout.

Download adapter source Download config example Get the portable AI bundle

From a checkout, run the protocol smoke test with node scripts/check-mcp.mjs. Launching the adapter directly with node assets/ai/mcp-server.mjs waits for an MCP client on stdin; silence is expected. Only protocol messages are written to stdout.

Tools	Behavior
`get_brand_context`	Read current facts, proposed standards, source priority and approval status.
`search_assets`	Search by words, category and format; limit results to 1–25 records.
`get_asset`	Read one exact resource ID, including published download links.
`list_asset_categories`	Discover the current categories and counts.

Fixed readable resources are andhub://brand-context, andhub://brand-data, andhub://brand-spec, andhub://guidelines and andhub://catalog. The optional andhub_brand_brief prompt combines a caller-supplied task with current context.

Local adapter, static hosting. GitHub Pages serves the source and context files. It does not run this adapter or provide an HTTP MCP endpoint. Your client must support launching a local stdio server; cloud-only clients may need the URL or attachment approach instead.

Protocol references: MCP transports, initialization, resources and tools.

Give your toolsthe right context.