Simular

Overview

Simular AI builds computer use agents that automate desktop tasks by interacting with your screen the way a human would. These agents click buttons, fill forms, navigate applications, and execute multi-step workflows across macOS and Windows without needing APIs or integrations. Founded by ex-Google DeepMind engineers, Simular targets industries with heavy repetitive computer work, particularly those without APIs for programmatic access. Insurance companies processing claims forms, healthcare organizations handling patient data entry, and back-office teams managing manual workflows are primary use cases. The platform includes a Workflow Editor for building and sharing automations, a community gallery for reusable workflows, and Simulang, a custom language for defining agent behaviors. With 10,000+ GitHub stars and a growing Discord community, Simular has built a developer ecosystem around computer use automation.

Features

• Computer vision control -- AI agents see your screen and interact with any application like a human user
• Cross-platform support -- Works on macOS (Apple silicon) and Windows for desktop automation
• No API required -- Automates applications that lack APIs by using visual interaction instead
• Workflow Editor -- Visual tool for building, testing, and refining agent automation sequences
• Simulang language -- Custom scripting language for defining precise agent behaviors and conditions
• Community gallery -- Share and discover reusable workflows built by other Simular users
• Full action logging -- Every click, keystroke, and decision is recorded for audit and review
• Virtual machine support -- Run agents on dedicated VMs for 24/7 autonomous operation
• Multi-step automation -- Chain complex sequences across multiple applications in a single workflow
• Team workflow sharing -- Share private automations within your organization with role-based access
• Browser automation -- Navigate websites, fill forms, extract data from web applications
• Developer-friendly -- 10K+ GitHub stars, Discord community, extensible through custom modules

Best For

Insurance and healthcare teams automating claims processing and patient data entry in legacy systems, Back-office operations filling forms and transferring data between applications that lack API access, Developers building computer use automations who want an extensible, community-driven platform, Enterprise teams replacing repetitive manual computer work with supervised AI agents

How It Works

Install the Simular agent on your macOS (Apple silicon) or Windows machine. The agent observes your screen and can control mouse, keyboard, and application interactions. You define tasks through the Workflow Editor or describe them in natural language. The agent uses vision-language models to understand what is on screen, make decisions about what to click or type, and execute multi-step sequences. For example, it can open a browser, navigate to a form, extract data from one application, and enter it into another, all autonomously. Every action is logged and visible. You can review what the agent did, audit its decisions, and refine workflows. Shared workflows can be published to the community gallery or kept private within your team. Virtual machine support on the Pro tier lets agents run on dedicated infrastructure for 24/7 operation.