AI Resources for Mambu Docs

This page lists all the AI-optimized documentation resources available from the Mambu Documentation Hub and explains how to fetch and use them.

Available Resources

All resources are served from https://docs.mambu.com and are publicly accessible.

Resource	URL	Description	Size
Full Export	/docs-export.json	Complete documentation export — all content, metadata, and structure	~140 MB
RAG Chunks	/docs-rag-chunks.json	Pre-chunked documents ready for vector database ingestion	~15 MB
AI Manifest	/ai-manifest.json	Navigation index listing all documents with titles and URLs	~2.5 MB
AI Sitemap	/ai-sitemap.txt	Plain-text sitemap with URLs, titles, and categories	Small
LLM Index	/llm.txt	Concise AI-optimized overview of the documentation hub	Small
XML Sitemap	/sitemap.xml	Standard sitemap for crawlers and search engines	Small

Fetching Resources

Full Export (recommended for RAG / AI agents)

# Download the full export
curl -O https://docs.mambu.com/docs-export.json

# Download compressed (faster — decompress on the fly)
curl -H "Accept-Encoding: gzip" https://docs.mambu.com/docs-export.json | gunzip > docs-export.json

RAG Chunks (pre-split for vector DBs)

curl -O https://docs.mambu.com/docs-rag-chunks.json

Chunks are configured with:

• Target chunk size: 1,000 tokens
• Overlap: 200 tokens
• Minimum chunk size: 100 tokens

AI Manifest (index / navigation)

curl -O https://docs.mambu.com/ai-manifest.json

Use this to list all available documents, browse by category, or build a table of contents without downloading the full export.

LLM Index (quick orientation)

curl https://docs.mambu.com/llm.txt

A compact summary of the docs hub — good for priming a model with context about Mambu before a conversation.

Resource Contents

docs-export.json

{
  "meta": {
    "version": "1.0",
    "exportDate": "2026-03-23T11:35:34Z",
    "source": "Mambu Documentation Hub",
    "baseUrl": "https://docs.mambu.com"
  },
  "statistics": {
    "totalDocuments": 3272,
    "totalWords": 467312,
    "totalTags": 533
  },
  "documents": [
    {
      "id": "docs/deposits/deposit-accounts",
      "title": "Deposit Accounts",
      "url": "https://docs.mambu.com/docs/deposits/deposit-accounts",
      "content": "Full markdown content...",
      "cleanContent": "Cleaned text optimised for AI...",
      "metadata": {
        "tags": ["deposits", "accounts"],
        "lastModified": "2026-01-09T12:52:40Z",
        "fileSize": 15420
      }
    }
  ],
  "documentsByCategory": { "api": [...], "docs": [...] }
}

docs-rag-chunks.json

{
  "meta": { "chunkConfig": { "targetSize": 1000, "overlap": 200 } },
  "statistics": {
    "totalChunks": 9963,
    "totalDocuments": 3365
  },
  "chunks": [
    {
      "id": "docs/deposits/deposit-accounts#chunk-0",
      "text": "Chunk text ready for embedding...",
      "metadata": { "title": "...", "url": "...", "tags": [...] }
    }
  ]
}

ai-manifest.json

{
  "overview": {
    "totalItems": 3388,
    "totalDocuments": 3272,
    "totalApiSpecs": 116,
    "categories": 10
  },
  "categories": {
    "api": { "count": 2901, "items": [{ "id": "...", "title": "...", "url": "..." }] },
    "docs": { "count": 371, "items": [...] }
  }
}

Usage Examples

Load into a Python AI agent

import json
import urllib.request

# Download the export
url = "https://docs.mambu.com/docs-export.json"
with urllib.request.urlopen(url) as response:
    docs = json.loads(response.read())

print(f"Loaded {docs['statistics']['totalDocuments']} documents")

# Filter by tag
deposit_docs = [
    d for d in docs["documents"]
    if "deposits" in d["metadata"].get("tags", [])
]

Ingest RAG chunks into a vector database

import json
import urllib.request

url = "https://docs.mambu.com/docs-rag-chunks.json"
with urllib.request.urlopen(url) as response:
    data = json.loads(response.read())

for chunk in data["chunks"]:
    vector_db.upsert(
        id=chunk["id"],
        text=chunk["text"],
        metadata=chunk["metadata"]
    )

Browse available documents (JavaScript)

const response = await fetch("https://docs.mambu.com/ai-manifest.json");
const manifest = await response.json();

// List all API docs
const apiDocs = manifest.categories.api.items;
console.log(`${apiDocs.length} API documents available`);

Document Categories

Category	Description	Doc Count
api	API v1, v2, Streaming, and Payments reference	~2,900
docs	User guide and conceptual documentation	~370
documentation-guidelines	Internal documentation standards	small

Regenerating the Resources

Resources are regenerated automatically as part of the site build (postbuild). To regenerate manually from the repo:

# Regenerate all AI resources
yarn ai:all

# Or individually
yarn export-docs          # docs-export.json
yarn ai:manifest          # static/ai-manifest.json
yarn ai:rag               # static/docs-rag-chunks.json
yarn ai:sitemap           # static/ai-sitemap.txt

The scripts live in scripts/:

• scripts/export-docs-for-ai.js
• scripts/generate-ai-manifest.js
• scripts/generate-rag-chunks.js
• scripts/generate-ai-sitemap.js

Related Docs

• Documentation Export Tool — full reference for the export script
• AI Export Quick Start — end-to-end example with LangChain and vector DBs