I always have so much fun building these text-to-speech chatbots, especially using ElevenLabs API (120 voices) compared to OpenAI (only 6 voices so far). This simple test chatbot feels like I'm in a 'Mission: Impossible' or 'Jason Bourne' movie, so I have to publish it so others can try also. https://mediafiles.botpress.cloud/f121b51d-93f7-4ef4-a98f-ec86a7a437c0/webchat/bot.html Try usernames: janesmith, johndoe, carlosgarcia. Then ask "What is my task?" Original chatbot: Automatically create private Knowledge Bases for users or user groups. The central Knowledge Base contains all the information, but chatbot users do not have direct access to it. When the user initiates a chat with the bot, it first requests their username. The bot then creates a private Knowledge Base containing only the information that the current user is permitted to access, and uses this to provide answers to their questions. When a new user logs in, the chatbot rebuilds the private Knowledge Base with the information that this new user is allowed to access. https://discord.com/channels/1108396290624213082/1195763907680280637 I'm changing all my old chatbots here to use text-to-speech so that I really get the gist of how everything works and how it's best to build these. If I get the ElevenLabs grant (not for these small chatbots, but for my flagship project), I'll share that usage (free 11,000,000 characters, 200 hours per month, for 3 months) with everyone who has also built text-to-speech chatbots (using my versions, Decay's, their own version, or some other version). Because we need to TEST these first so much that everything works 100% before offering these to our clients and users. So when we have first built enough of these, I'll ask ElevenLabs to check these also (not only the project I'll send them) if we are worth it 🦾 🛠️ 💰