Meta has confirmed that Manus, the Singapore-based AI agent company, is joining its growing AI portfolio. On paper, the deal brings a fast-scaling general-purpose agent into Meta’s ecosystem. In practice, it pulls a controversial but widely used product into one of the biggest technology companies on the planet.
Manus has spent the past year positioning itself as an “execution layer” for AI. Instead of just answering questions, its agent takes on work: research tasks, automation workflows, and end-to-end projects that blend code, content, and data. The company claims its agent has already processed more than 147 trillion tokens and spun up more than 80 million “virtual computers” to carry out tasks. Those numbers raised eyebrows, and not just in a good way.
From Monica to Manus: How the Agent Came Together
Manus did not appear out of thin air. The parent company, Butterfly Effect Technology, sits on a track record of shipping productivity products. Its previous tools, such as AI-enabled assistants for Chinese business users, reached millions and attracted investment from major regional players. That user base and infrastructure set the stage for Monica, a browser extension that integrated multiple large language models.
Monica gave the team two key assets: real users and clear telemetry on how people actually work with AI. Manus built on that foundation. The team took lessons from Monica—long-form research, task breakdown, and cross-site workflows—and turned them into the core Manus agent. By early 2025, Manus launched publicly, raised a large funding round, and reported annual recurring revenue above $100 million in its first year, according to references shared on Hacker News.
What Manus Actually Does for Users
Under the hood, Manus connects to commercial models such as Claude, and mixes them with tools that can browse, code, and handle documents. Instead of a simple chat window, each “conversation” can run inside its own isolated environment that behaves like a lightweight Linux machine. Manus can write code, run it, debug it, collect results, and then loop back to refine its work.
The agent supports structured outputs as well. Users can ask for tables, long-form markdown reports, PDFs, slide decks, or extracted data. This framing has resonated with people who want to turn text into action rather than yet another paragraph of prose.
Why Digital Marketers Care About Manus
Manus has also picked up a dedicated following in digital marketing circles. Marketers have used it heavily to produce long-form content, cluster pages, and landing copy at scale. Others lean on it to generate information architecture, content briefs, and draft web pages that feel more original than the usual template output.
Some practitioners say Manus produces stronger marketing copy and more unique website structures than many popular chat-style tools. They point to workflows where Manus performs deep research on a topic, outlines an entire site, drafts SEO (Search Engine Optimization) content, and even proposes internal linking strategies. For people juggling campaigns, ad copy, and client reports, the pitch is simple: less prompting, more finished work.
Meta’s Strategic Angle: Agents, Users, and Distribution
Meta already has strong model assets through its Llama family. What it has lacked is a standout consumer-grade agent that people outside its apps talk about daily. Manus fills that gap. It brings an existing subscription business, millions of users, and a product that already functions as a general-purpose agent capable of research, coding, and automation.
Meta’s own announcement emphasizes a clear plan: continue to sell the Manus service directly, integrate its agent into Meta AI, and extend the agent to businesses and individuals across Meta’s platforms. If that happens, Manus could become a default “do the work for me” layer inside products that billions already use, from messaging to productivity.
Hacker News Reaction: Admiration, Skepticism, and a Lot of Questions
The Hacker News thread on the acquisition reads like a real-time debate panel. The sentiment is mixed. You see genuine respect for Manus as a product. You also see sharp criticism of the valuation, the business model, and Meta’s involvement.
What Commenters Like About Manus
Several developers and founders describe Manus as one of the best agents they have used for turning text into completed work. They note that Manus can take a prompt and produce usable slide decks, working code, and structured research outputs with less babysitting than many alternatives. One commenter mentions Manus as “the best agent for turning text into work” and highlights its ability to handle tasks like data extraction and code-driven workflows with minimal configuration.
Others compare Manus favorably to tools from the United States. In their view, Manus and competitors like Kortix stand out because each “chat” runs in its own environment, making the agent feel more like a flexible worker than a simple chatbot. Manus’s early support for deep research behaviors also impressed people before similar features became standard in offerings from OpenAI and others.
Concerns Over Valuation, Hype, and Margins
That praise sits next to heavy skepticism. Many commenters look at the rumored revenue numbers and token usage and see a business that might be burning cash at an alarming rate. If Manus processed 147 trillion tokens and sits near $100 million in ARR, several participants argue that API costs could be enormous, especially if premium models are involved. One commenter speculates that Manus may have needed an acquisition simply to keep the lights on.
Another theme is the broader AI funding climate. Commenters talk about “dumb money” chasing any AI story with momentum. Some frame the Manus deal as part of a wider pattern where founders build an attractive wrapper around frontier models, generate hype through influencers, then exit quickly before economic gravity kicks in. References to bubbles, money laundering analogies, and comparisons to the dot-com era appear throughout the thread.
Debate Around Product Depth and “Wrapper” Critiques
Several voices describe Manus as “just a wrapper around Claude” with strong marketing. For them, the acquisition feels like Meta paying a premium for API calls, not for deep proprietary technology. Others push back, pointing out Manus’s agent flows, browser automation, sandboxed “virtual computers,” and context engineering work. They argue that if it were pure hype, competing clones would have matched it more convincingly by now.
This disagreement reflects a broader split in the AI community: some people see agents as thin glue code; others see orchestration, reliability, and UX as the real moat. Manus sits squarely in that debate.
National Origin and Stereotypes
The thread also includes a heated side discussion about Manus’s Chinese roots and later move to Singapore. One founder from China calls out what they see as unfair stereotyping of Chinese startups, arguing that Manus had a strong reputation, a profitable predecessor product, and serious investors. Another Chinese founder counters that Manus leaned heavily on marketing and that being an AI chatbot aggregator is not technically impressive on its own.
The exchange exposes a familiar pattern: some observers are quick to label Chinese-origin startups as hype-driven or politically connected, while others stress that the quality and ambition vary just as much as in any other ecosystem.
Meta’s Track Record and Trust Issues
Meta’s role adds another layer of friction. Many commenters openly distrust the company, referencing its history with social platforms, news feeds, and the metaverse effort. They question whether Meta will keep Manus innovative or slowly fold it into a large machine where product quality takes a back seat to internal politics and ad-driven incentives.
Several users say they stopped using Manus earlier over privacy concerns, only to joke that Meta’s involvement does not ease those worries. Others worry about talent retention. The pattern with past acquisitions—Oculus, for example—haunts this conversation. People ask if Manus will remain distinctive or fade into a generic “Meta AI” feature over time.
How This Looks From a Product and Strategy Lens
Stepping back, Manus sits at an interesting intersection. It is one of the first high-profile agents built primarily as an execution tool for work. It has already been adopted by engineers, researchers, and a growing wave of digital marketers who use it to generate content, build sites, and ship campaigns. It has proven demand at scale, even if the unit economics remain a question mark.
For Meta, the acquisition delivers three assets at once: a proven agent UX, a paying customer base, and a team that has already wrestled with the hard parts of orchestration and context management. For Manus’s investors and founders, it offers a fast path to liquidity in a market where many AI startups will never see an IPO window.
For users, the story is more personal. If Manus is part of your daily workflow—whether that means drafting content, coding, or building client websites—you now live with a new kind of platform risk. The company promises that the service will continue to run as a subscription product, with operations staying in Singapore. Many Hacker News readers say they have heard similar assurances before and still watched products lose their edge after being absorbed by giants.
In the end, the Manus–Meta deal captures this AI moment better than any marketing line. There is real product value here. There is real hype here. There are very real questions about cost, control, and trust. Developers, marketers, and founders see the same facts and come away with very different conclusions. That split will only grow sharper as more AI agents, more acquisitions, and more billion-dollar bets hit the news over the next year.