Town

Town is an AI work assistant whose agent, called a Townie, works across email, calendars, chat and business apps. It learns your people, projects and writing voice, then drafts, summarizes and organizes work with you, with approval before actions.

Town is an AI work assistant built around a personal agent called a Townie. Instead of living in a single chat window, the Townie operates across the places where work actually accumulates: inboxes, calendars, chat tools and business apps. Over time it builds a working profile of the person it serves, covering the people they deal with, the projects underway, their goals, how their week is structured and the tone they use in email. That profile is visible in the product as pages such as Overview, People, Projects, Goals and How you work, and it is updated as new activity comes in. The practical pitch is that the usual setup step of explaining yourself to an assistant disappears. Example prompts on the homepage show the pattern: a Townie drafts a reply to a client who wants to reschedule using open slots from the calendar, summarizes a busy Slack channel into the questions still waiting for an answer, logs calls in a CRM, matches receipts to card charges, or groups bug reports into issues. Actions that change something, such as moving a meeting or buying an item, are described as waiting for approval first. The site lists integrations across messaging, calendar, meetings, CRM, file storage, project management, analytics, development and finance, and it also mentions support for any MCP server. A Townie is also reachable by text message and from an iPhone app. Town targets individual professionals and small teams with heavy coordination loads, such as business owners, recruiters, sales staff, investors and solo practitioners running client work. The homepage sample profile, a solo interior designer who needs faster proposals and less evening invoicing, shows the intended audience well. On trust, the company states that nothing a user connects trains a model by default and that it is SOC 2 Type 2 compliant, with security and subprocessor pages published. Among alternatives, Town sits between general chat assistants that need constant context and narrow single-app automation tools. Its distinguishing bet is persistent, cross-app memory of the user's working life, which suits people whose work is scattered across many tools more than those who live in one.