Back to directory
AI & ML · AI Agents

Manus

general AI agent that bridges minds and actions: it doesn't just think, it delivers results.

Manus from @Meta is the general AI agent that bridges minds and actions: it doesn't just think, it delivers results. Telegram: https://t.co/kdHdNxZ6xF
Singapore249K followers
TLVC Rating

The video was done well, but the production could have been better.

Community Rating
No ratings yet
Your rating
Sign in to rate this launch.

About

Manus Cloud Computer gives the company's general AI agent a persistent virtual machine in the cloud, so tasks and applications kept by the agent continue running even when the user closes their laptop or puts down their phone. Behind Manus is a multi-agent system that executes real-world workflows end to end, with prompts going first to a planner agent that decomposes them , and each Sandbox is a fully isolated cloud virtual machine that Manus allocates per task, running in its own environment with networking, a command line, a file system, and a browser . The launch matters because Manus has been positioning itself less as a chat interface and more as an operating environment, and an always-on cloud machine is the piece that lets agent-built workflows behave like real software rather than one-off sessions. The product is aimed at builders and operators who want an agent to handle multi-step work without babysitting a local process. In Manus 1.5, full-stack web application development became a major capability, letting users build and deploy production-ready web apps through conversation, with the integrated browser actively testing what it builds . With Cloud Computer, those generated apps and long-running jobs can stay live around the clock, accessible from the Manus mobile app or web version, so users can start on a laptop, make updates from a phone, and deploy changes from a tablet . Manus is built by Chinese startup Butterfly Effect, founded by CEO Red Xiao Hong, who has described the agent as more like a human being than a chatbot , and chief scientist Yichao "Peak" Ji, the 32-year-old who introduced the product in its launch video . The underlying system runs on existing LLMs including Anthropic's Claude and fine-tuned versions of Alibaba's Qwen , and the team has been iterating quickly on features like wide research, desktop integration, and now a permanent cloud runtime for whatever its users decide to build.
Tags
PublicB2C<500KAI agentProduct launchB2BDemoUS
Comments (7)
Sign in to join the discussion.
Rohan Vaidya4/30/2026

Bio says Manus is from Meta now? Either someone fat-fingered the launch copy or there's a reveal I missed.

Lena W.4/30/2026

The tweet buries the lede. 'Always on machine in the cloud' should have been the hook, not paragraph two.

Priscilla Ng4/30/2026

Tested the mobile flow. Three taps to a running agent is genuinely tight, but the empty state on first launch is a wasteland.

kweku4/30/2026

Anyone can build, anything can run, my electricity bill can also do whatever it wants apparently.

Tomasz K.4/30/2026

Cloud sandbox for agents is the obvious next step, but the real question is what happens when my agent racks up a 14 hour loop overnight.

Amara Okafor4/30/2026

Launch video cuts are clean but the voiceover sounds like it was recorded inside a Tupperware. Someone get this team a pop filter.

Vik Bhatla4/30/2026

Built something adjacent and the 24/7 framing is smart positioning. Curious how you're handling state when a session crashes mid-task.