A multi-provider AI chat assistant. Switch between Groq, Mistral, Gemini, OpenAI, Claude, DeepSeek, or Azure OpenAI without touching the frontend, ground its answers in your own FAQ or documentation with a CSV drop, and give it a defined persona instead of a generic AI voice. Built with encrypted memory, rate limiting, and real security from the ground up. Try it in the corner.
Groq, Mistral, Gemini, OpenAI, Claude, DeepSeek, or Azure OpenAI, switch between them from a dropdown, no code changes.
Conversation history is AES-256-GCM encrypted at rest in MySQL, not stored as plain text.
Tied to a visitor cookie, history survives page reloads and return visits, and auto-expires after 30 days of inactivity.
Built-in abuse protection by IP and by visitor, checked before any paid API call ever goes out.
Bold, links, code, lists, and more, rendered safely. HTML is escaped first, so nothing an AI replies with can ever inject a real tag.
Drop a CSV of FAQs, policies, or documentation, and answers get grounded in it automatically. Not a generic chatbot, one that actually knows your business.
Give it a defined role, a sales associate, a coding tutor, a support agent, instead of a generic AI. Off by default, and hardened against "ignore your instructions" attempts when it's on.
One include line embeds it on any PHP page, anywhere on your site, see below.