Early access: connect a computer your company owns, and an AI employee answers with an AI model installed on it. Its memory, files and chat history stay in VEP's cloud.
The AI model on your computer writes the employee's main answers. The employee's memory, files, chat history and logs stay in VEP's cloud, which builds every request it sends to your computer and keeps every answer. VEP's cloud also uses cloud AI models for some replies and steps, for example in chats with your customers: short replies to greetings and a running summary of each conversation.
A Linux computer with an NVIDIA GPU. You install the VEP software with terminal commands, connect it with a code from VEP and download an open model to it. The setup guide shows every step.
Local inference costs €0 per token: VEP does not charge for answers from your computer, and they do not use up an AI employee's daily spending limit. You pay for the computer and its power. Replies to your customers still count toward your plan's monthly message limit.
How fast an answer comes depends on your computer and the AI model you choose. Every request still travels from VEP's cloud to your computer and back.
An AI employee you place on your computer gives its main answers only while that computer is online. While it is offline, no cloud AI model writes those answers in its place. In chats with your customers, some replies, for example to a greeting, can still come from VEP's cloud. Your other AI employees keep working in VEP's cloud.
VEP switches your own hardware on for selected companies. Ask us to switch it on for yours.
How much of an AI worker can sit on one computer in your own office: the memory alone, the memory and the model, or all of it. Three placements explained in plain words, with animations and specs. Today the model can run there in beta; the rest is on request.
Your own hardware is in early access. Setup takes a Linux computer with an NVIDIA GPU, a few terminal commands and an AI model download.