Perxona Connect Kit

Studio Demo

Presenter Setup

Avatar

Scene

Voice

Model source

Who answers. Switch any time — the conversation carries over, so you can ask the same question of both.

Connect Chatbot Perxona runs the model. You get the chatbot below: custom instructions, knowledge files and function tools, all stored server-side. Your own LLM You run the model, with your LLM_API_KEY. No chatbot to manage — and no knowledge files or tools either, since those live with the chatbot.

LLM_API_KEY is not set, so /api/chat returns 501. Set it in .env and restart to enable this option.

Chatbot

Active Chatbot

Create / Edit

Name * System Instructions

Knowledge File No file

Upload a .txt, .pdf, .doc, .docx, or .csv file. The chatbot will use its content to answer questions (RAG).

Choose file…

Function Tools

JSON array of HTTP APIs the chatbot can call. Leave empty to keep existing tools unchanged. Use [] to clear all tools.

Select an avatar & scene,
then click Launch Presenter.

Select a chatbot and launch the presenter to start chatting.