David One

David One is an AI personal assistant that combines web search with cited answers, document and media analysis (audio, video, images), cross-chat memory, group chats and scheduled tasks. Currently in free beta, it targets individuals and small teams who want research and routine work handled in one shared workspace.

David One is a web-based AI assistant built around a simple idea: answer questions and finish tasks inside one conversation, with evidence attached. It pairs a chat interface with live web search and browsing, so responses can draw on current information rather than a static training window. Each answer comes with cite notes pointing back to the sources, which makes it easier to check a claim before acting on it. Beyond text chat, David handles files and media. Users upload documents and ask for specific information across languages, with citations at the page and paragraph level. The assistant also works with audio (meeting notes, summaries, targeted questions) and video (summaries, locating a specific scene by time), plus screenshots, charts and photos for reading text or extracting data. A long term memory layer, described as brain-inspired, carries context between sessions so repeat questions need less setup. Productivity features push it past a standard chatbot. Group chats let several people work with David in one shared space, useful for small teams coordinating research or drafting. Scheduled tasks run in the background on a recurring basis, and a built in todo list with task breakdown, reminders and notes covers everyday routine management. The beta is open at no cost, with an enterprise plan in early access, though specific tiers and limits are not published on the homepage. David One competes with general assistants such as ChatGPT and Claude and with research tools that emphasize source citation. Its differentiators are the combination of granular file citations, group collaboration and media understanding in a single product, plus the memory system. Buyers who need documented audit trails or team-shared AI workspaces will find the model worth evaluating, while anyone needing published pricing or third party integrations should wait for more detail.