Factory brings its “Droids” software development agents out of the terminal with new desktop app

AI coding agents have largely lived in terminals, IDE plugins, or browser tabs. Factory, a three-year-old startup building AI agents for software development, reckons that isn’t enough.

The company this week introduced a native desktop application for its “Droids,” a system of AI agents designed to handle software development tasks such as debugging, migrations, and code generation.

Factory debuted Droids back in September, alongside a fresh $50 million in funding, positioning the system as a way to automate complex engineering workflows. Its new app, available on macOS and Windows, provides a dedicated interface for running and managing those agents across multiple workflows.

Factory’s desktop app: A quick peek

Before the desktop app, Factory’s Droids were primarily accessed through the command line. That setup suited engineers comfortable scripting and orchestrating agents directly, but it also limited how widely the tools could be used across teams.

The desktop app changes that by introducing a persistent, visual layer on top of those same capabilities.

Users can run and monitor multiple “Droid” sessions, each maintaining its own context and history, with tasks unfolding step by step and remaining accessible as work progresses.

A key addition is what Factory calls “Droid Computers.” Instead of starting fresh each time, agents operate on persistent machines that retain installed packages, repositories, and running services between sessions. These environments can be provisioned in the cloud or connected to a user’s own hardware, including systems running local models for more controlled deployments.

The app also introduces new “computer use” functionality, expanding how agents interact with software.

So rather than operating in a confined prompt-response loop, Droids can control applications directly—navigating development environments, interacting with browser tabs, or executing commands in a terminal.

For example, a Droid can open a spreadsheet, extract and analyze data, generate a pivot table, and apply formatting, all while tracking each step of the process. The idea is to move beyond text-based assistance toward agents that carry out tasks across the full computing environment.

Agents become persistent

Factory’s move mirrors a broader trend. As AI agents take on longer-running and more complex work, companies are moving away from purely text interfaces toward tools designed to supervise and coordinate them.

A similar direction is visible in Anthropic’s recent updates to Claude, where features like Dispatch and Cowork introduce persistent sessions that can run in the background on a user’s machine. In both cases, the emphasis is on agents that don’t just respond to prompts, but continue working over time—handling tasks across files, tools, and applications without constant input.

There are also parallels in how those agents interact with local environments. Anthropic’s computer use feature allows Claude to operate a mouse and keyboard, navigating software directly with user approval. Factory’s approach follows a comparable path, giving Droids the ability to move across applications and systems as part of a single task, rather than relying solely on APIs or predefined integrations.

For Factory, the desktop app is perhaps less about adding new capabilities than it is about making existing ones easier to manage. The underlying system – agents that can be triggered from terminals, IDEs, or APIs – remains the same. What changes is how those agents are observed, controlled, and extended across an organization.