Manus

The general AI Agent: it autonomously completes real tasks — research, website building, development, slide decks, and more — delivering finished products, not just suggestions.

Built by the Butterfly Effect team, Manus burst onto the scene in March 2025 as “the world’s first general AI Agent,” with a core pitch: taking a task from start to finish. Give it a single instruction, and it will plan on its own, call on a browser and code environment, and deliver a finished research report, website, app, or slide deck. In late December 2025, Meta announced it had acquired Manus for more than $2 billion — the official site now displays “Manus is now part of Meta,” but the subscription service continues to operate independently.

The current flagship is the Manus 1.6 series, available in three tiers: Lite, Standard, and Max. Paid users can run up to 20 concurrent tasks, and the platform pairs Wide Research for parallel research, Design View as a visual canvas, and Cloud Computer as a computer in the cloud — covering end-to-end scenarios that range from mobile app development to brand design.

The update cadence over the past three months has been remarkably intense: July’s Branch lets the same context fork into multiple parallel directions; June brought managed mode and the Zoom connector; May added Scheduled Tasks 2.0 and the Shopify connector. The connector ecosystem is fast becoming the moat that sets Manus apart from pure conversational products.

Pricing is credit-based: the free tier offers 300 credits a day to try things out, while paid plans start at $20/month and unlock the full model lineup along with concurrency. It’s best suited to heavy users willing to pay for work done on their behalf; for lightweight Q&A needs, it may not offer the best value.

Source: Agentopia · See /llms.txt for the machine-readable catalog of the entire site