AI integration

Use the built-in AI assistant, load the docs into your AI editor, or connect any LLM directly to the BMONI Embedded documentation.

BMONI Embedded docs ship with several AI integrations out of the box. You can chat with the docs, open any page directly in your AI editor, or point an MCP-compatible agent at the full documentation corpus.

---

Built-in AI assistant

Click the Ask AI button in the top-right of any page to open the Mintlify assistant. It is trained on all BMONI Embedded documentation — SDK, UI Kit, Wallets & Cards, and the API reference — and can answer questions, generate integration code, and walk you through setup flows.

The assistant is always up-to-date with the latest published docs. No extra configuration is needed.

---

Open any page in your AI editor

Every page has a contextual toolbar (bottom-right of each code block) with one-click actions for:

Cursor

Opens the page in Cursor’s AI chat with full context pre-loaded.

Claude

Opens the page in Claude.ai as a new conversation.

ChatGPT

Opens the page in ChatGPT with the content pre-filled.

Windsurf

Opens in Windsurf IDE with context injected.

Google AI Studio

Opens the page in AI Studio.

Perplexity

Opens as a Perplexity research query.

---

MCP — connect any AI agent to the docs

Mintlify exposes the full documentation corpus as an MCP (Model Context Protocol) server. Any MCP-compatible agent — Claude Desktop, Cursor, Windsurf, Devin — can query the docs as a tool.

Add to Claude Desktop

1. Open Claude Desktop settings

   Go to Settings → Developer → Edit Config (claude_desktop_config.json).

2. Add the BMONI docs MCP server

   claude_desktop_config.json

   {
     "mcpServers": {
       "bmoni-embedded-docs": {
         "command": "npx",
         "args": [
           "-y",
           "@mintlify/mcp@latest",
           "--docs-url",
           "https://embedded-dev.bmoni.com/docs"
         ]
       }
     }
   }
   
3. Restart Claude Desktop

   Quit and reopen Claude Desktop. The bmoni-embedded-docs tool will appear in the tool list.

4. Start asking

   Ask Claude anything about the SDK, UI Kit, or wallet cards — it can now retrieve exact content from the docs.

   “Show me how to provision a wallet with bmoni_embedded_sdk and gate signing with a PIN.”

Add to Cursor

Open Cursor Settings → MCP and add:

{
  "bmoni-embedded-docs": {
    "command": "npx",
    "args": [
      "-y",
      "@mintlify/mcp@latest",
      "--docs-url",
      "https://embedded-dev.bmoni.com/docs"
    ]
  }
}

Cursor’s AI will now retrieve BMONI documentation when you ask about the SDK, UI Kit, or wallet card components.

---

llms.txt — raw documentation index for LLMs

Mintlify auto-generates two plain-text files that any LLM can fetch directly:

URL	Contents
/llms.txt	Page titles and URLs — a lightweight index for agents to locate content
/llms-full.txt	Full page content for every page — paste into any LLM context window

# Fetch the lightweight index
curl https://embedded-dev.bmoni.com/llms.txt

# Fetch the full content dump
curl https://embedded-dev.bmoni.com/llms-full.txt

Use in a system prompt

If you are building an app that wraps an LLM, you can include the full docs as context:

const docsContent = await fetch(
  'https://embedded-dev.bmoni.com/llms-full.txt'
).then(r => r.text());

const response = await openai.chat.completions.create({
  model: 'gpt-4o',
  messages: [
    {
      role: 'system',
      content: `You are a BMONI Embedded integration assistant. Here is the full documentation:\n\n${docsContent}`
    },
    { role: 'user', content: userQuestion }
  ]
});

---

Quick-copy for AI prompts

Every code block in the docs has a Copy button and an Ask AI shortcut. For setup questions, the most effective prompt is:

Using the bmoni_embedded_sdk Flutter package, I need to:
1. Provision an on-device Ethereum wallet
2. Gate signing with a 6-digit PIN
3. Sign an EIP-191 message and return the signature

Show me the full Dart implementation.

Paste this into any of the AI editors above — the contextual toolbar pre-fills the relevant page content automatically.