Inception Chat

Inception Chat is the web chat interface for Mercury 3, a diffusion LLM from Inception. It offers a simple prompt box, suggested prompts geared toward coding and explainers, and chat history that is saved locally in the browser rather than on the company's servers.

Inception Chat is the browser-based chat interface for Mercury 3, which its maker describes as the fastest commercial-grade diffusion LLM. The experience is deliberately plain: a prompt box, a set of suggested prompts and a local chat history. There is no workspace scaffolding, no file library and no agent builder. What the product sells is the model underneath, and the interface is designed to get out of its way so a prompt goes in and an answer comes back quickly. The suggested prompts reveal the intended audience. Most of them are build requests: Flappy Bird in HTML5 and JavaScript, a Kanban board with drag and drop, a regex playground with live capture-group highlighting, a pathfinding visualizer for BFS, Dijkstra and A*, Conway's Game of Life. Others lean toward explanation and writing, such as a Fourier series explainer with rotating epicycles, special relativity described for a five year old, a product blurb translated into seven languages with back-translation, or monthly expenses turned into a budget dashboard. In practice this is a general assistant that leans toward developers, students and analysts who want working code snippets or interactive explainers rather than polished long-form marketing copy. Working with it is straightforward. Open the site, type a prompt or pick a suggested one, read the response and continue the thread. Chat history is stored in the browser through local storage rather than on Inception's servers, so conversations stay private to that machine but do not sync between devices or browsers, and clearing site data will remove them. That design choice makes the tool pleasant for quick experiments and awkward for anyone who wants a persistent archive. The competitive picture is a crowded one. Inception Chat sits alongside general assistants from OpenAI, Anthropic, Google and others, but its differentiator is the diffusion-based Mercury 3 model rather than a feature set. Anyone who needs document handling, team spaces, plugins or mobile apps will find more complete products elsewhere; anyone who wants a fast, low-friction window into a specific model family will find this a clean way to try it.