Manus Launches Cue AI Agents, Expands Automation Features
Manus, a China-founded artificial intelligence startup now based in Singapore, has launched Manus 2.0, the next generation of its AI agent platform, alongside Cue, a new standalone app for deploying personal AI agents.
Manus 2.0 brings a new AI agent architecture, cloud computing capabilities, and creative tools, while Cue focuses on helping users delegate everyday tasks to AI agents that can send messages, make calls, and carry out tasks on users' behalf.
Key Features of Manus 2.0
Manus 2.0 introduces a new architecture for its AI agents, which reduces the time and cost involved in completing tasks. The update also expands what users can do with the platform, adding tools for building projects, automating workflows, and running tasks in the background.
At the core of Manus 2.0 is Cascade, the latest version of the company's in-house agent system. According to Manus, the new system loads tools only when needed and allows multiple actions to run simultaneously, helping it complete tasks more efficiently.
Manus claims that Cascade used 23.2 per cent fewer tokens, completed tasks 28.2 per cent faster, and cut costs by 32 per cent compared with the previous system.
Another key feature of Manus 2.0 is Cloud Computer, which will let users run projects even after they close their laptops. For example, users building multiplayer games will be able to host servers that stay online, while other projects and automated tasks can continue running in the background.
Manus is also expanding its automation features, allowing users to trigger workflows based on events in connected services instead of scheduling them for a specific time. For instance, a new email, calendar event, or Slack message can automatically trigger a task.
The company has also upgraded its desktop app to Manus Studio, bringing together tools for creating documents, spreadsheets, and presentations, building websites, and writing code. The workspace also includes dedicated tools for video editing and game development.
The new Video Editor will let users edit clips, images, text, motion graphics, and audio directly on a timeline. This means users can make changes to specific parts of a video without having to generate the entire video again. It also includes an Alchemy mode, which the company says can turn an idea into a video that users can further refine in the editor.
Meanwhile, the Game Dev tool will let users build and edit games using AI-generated code, images, and video, and publish playable versions online. The company says users will also be able to use Cloud Computer to set up multiplayer servers for their games.
Manus 2.0 also introduces Remote Control and Computer Use, which will let users direct the AI agent to work with files, browse the web, and interact with apps on their computers through an authorised session.
Introducing Cue, a Personal AI Agent App
Alongside Manus 2.0, the company has introduced Cue, a standalone app for personal AI agents that works across phones and computers. Built on the same infrastructure as Manus, Cue will let users assign tasks to AI agents, each with its own email address, phone number, digital wallet, and computer.
Users will also be able to bring multiple agents into a group chat, allowing them to divide tasks and work towards a shared goal. Cue is designed to work with everyday services, too. For example, an agent could scan a restaurant's QR code and place an order on the user's behalf.
Users will also be able to ask agents to send messages, make calls, and make payments within a set budget.
Manus 2.0 is available on the web, desktop, and mobile in early access. Cue is also being introduced as a standalone app, with access currently available through an invite code.
