Guides & Use Cases

Practical step-by-step guides for common MDDB use cases. Each guide walks through a real scenario from setup to deployment.

AI & Chat

AI Chat Widget for Any Website

Add a RAG-powered AI chatbot to any website. Runs entirely locally with Docker and Ollama โ€” no cloud API fees.

Stack: MDDB + mddb-chat (Rust) + mddb-chat-widget (TypeScript) + Ollama
Time: ~30 minutes
Cost: Free (local LLM)

โ†’ Read the Website Chat Guide

WordPress Content Analyzer

Export your WordPress site into MDDB, then use Claude or any LLM to analyze it. Find broken links, missing meta descriptions, duplicate content, and SEO gaps โ€” all from a single conversation.

Stack: MDDB + WordPress export + Claude CLI
Time: ~45 minutes
Use when: You have a WordPress site and want AI-assisted content auditing

โ†’ Read the WordPress Analyzer Guide

YouTube Channel Transcriber

Scrape transcripts from an entire YouTube channel, load them into MDDB, then query them with an LLM. Ask questions across hundreds of videos without watching them.

Stack: MDDB + yt-dlp + Claude CLI
Time: ~30 minutes
Use when: You want to search and analyze video transcripts with AI

โ†’ Read the YouTube Transcriber Guide

Common Patterns

AI Agent Memory

Store conversation sessions, recall context with semantic search, and summarize past interactions. MDDB provides 6 dedicated MCP tools for conversational memory.

curl -X POST http://localhost:11023/v1/add \ -d '{"collection":"memory","key":"session-abc","lang":"en_US", "meta":{"session":["abc"],"type":["conversation"]}, "contentMd":"User asked about billing..."}' curl -X POST http://localhost:11023/v1/vector-search \ -d '{"collection":"memory","query":"billing questions","topK":5}'

See MCP Server Config for the full memory MCP tools reference.

RAG Pipeline

Auto-embed documents on ingest, retrieve context via hybrid search, expose to LLMs through MCP.

curl -X POST http://localhost:11023/v1/ingest \ -F "[email protected]" \ -F 'meta={"collection":"kb","key":"docs","lang":"en_US"}' curl -X POST http://localhost:11023/v1/hybrid-search \ -d '{"collection":"kb","query":"how to configure auth","topK":5,"alpha":0.5}'

See RAG Pipeline for the complete guide.

Documentation Chatbot

Import documentation, embed automatically, expose via MCP โ€” instant AI-powered support for your docs.

  1. Upload your docs: POST /v1/upload (supports .md, .pdf, .docx, .html)
  2. Enable embeddings: set MDDB_EMBEDDING_PROVIDER=openai
  3. Connect via MCP: add MDDB to your Claude/Cursor/Windsurf config
  4. Ask questions directly in your AI client

See Chat Widget to embed a chatbot on your website.

Multi-Language Content

Store the same document in multiple languages using the same key, different lang field.

curl -X POST http://localhost:11023/v1/add \ -d '{"collection":"blog","key":"hello-world","lang":"en_US","contentMd":"# Hello World"}' curl -X POST http://localhost:11023/v1/add \ -d '{"collection":"blog","key":"hello-world","lang":"pl_PL","contentMd":"# Witaj ลšwiecie"}' curl -X POST http://localhost:11023/v1/get \ -d '{"collection":"blog","key":"hello-world","lang":"pl_PL"}'

More Examples

See the Examples page for code samples in curl, Python, PHP, and MCP configurations.