Quick start

Configure providers, ingest some content, and put the chat widget on your site.

1. Configure an LLM provider

ItihRag → Settings → LLM Providers — pick a provider, paste your key, choose a model. Click Test connection to verify.

2. Configure an embedding provider

ItihRag → Settings → Embeddings — do the same for embeddings. (Cloudflare’s free-tier Workers AI works for both.)

3. Choose a vector database

ItihRag → Settings → Vector Database — review the detected capability and pick an engine (or leave on Auto). If using Cloudflare Vectorize, click Create / verify index.

See Vector stores for the trade-offs.

4. Add content to your knowledge base

ItihRag → Knowledge Base — add a URL or upload a PDF, or use the WordPress Content tab to index your existing posts/pages.

5. Personalize the chat

ItihRag → Settings → Chat — confirm the bot name, welcome message and launcher position.

6. Visit your site

The chat bubble appears in the corner.

Embed the chatbot inline

Embed your chatbot anywhere with the shortcode:

[openrag_chat]

The floating widget can be enabled or disabled from Settings → Chat.

A fully local / offline setup

You can run the entire stack on your own hardware with no external requests:

  1. Install Ollama locally and pull a chat model (e.g. ollama pull llama3.1) and an embedding model (e.g. ollama pull nomic-embed-text).
  2. In Settings → LLM Providers, choose Ollama and point it at http://localhost:11434.
  3. In Settings → Embeddings, choose Ollama with the same base URL.
  4. In Settings → Vector Database, use MySQL (no external service).

A Cloudflare-only setup

Cloudflare Workers AI can serve both embeddings and chat, and Cloudflare Vectorize can host the vector index — one account for everything:

  1. In Settings → LLM Providers and Settings → Embeddings, choose Cloudflare Workers AI with your account ID and API token.
  2. In Settings → Vector Database, choose Cloudflare Vectorize and click Create / verify index.