Manus Launches Cue: AI Agent With Its Own Email, Phone Number and Wallet

Manus Launches Cue: AI Agent With Its Own Email, Phone Number and Wallet

Manus 2.0 was rolled out on September 28, 2026 by the Singapore-based company Butterfly Effect Pte. Ltd., together with Cue — a standalone app for personal AI agents that, according to the company, will each have their own email address, phone number, wallet and computer. The launch is the company's first major product release since it reportedly resumed fully independent operations on September 1, and comes alongside a reported $500 million funding round at a target valuation of around $4 billion. The performance figures behind the new Cascade architecture and the agent's payment and calling features are, for now, only the company's own claims, relayed through secondary sources.

What's new

Manus 2.0 is described by its developer as a wholesale redesign of the agent platform, not merely an update. "Manus 2.0 is not a version update... It's a new architecture," Butterfly Effect explains, according to Thurrott.com.

Alongside the platform comes Cue, a standalone app for dispatching personal AI agents, positioned as a consumer product rather than an enterprise one. It is currently in invite-code early access, which recipients can share onward, and can be tried for free. An iOS version is said to be coming soon, according to Thurrott's account of the company (Thurrott.com).

The date surrounding the launch is not entirely unambiguous: Cryptobriefing dates the rollout to September 28 (Cryptobriefing), while India Today and a Japanese review on note.com are dated September 29 (India Today; note.com). The difference may be due to time zones, but the exact announcement timing cannot be established from the available sources.

Cascade: the architecture meant to make agents cheaper

At the core of Manus 2.0 is Cascade, an agent architecture the company frames as a fundamental design choice rather than a performance tweak. According to Manus's official blog, as relayed by いしたにまさき on note.com, one test configuration produced 23.2 percent fewer tokens, 28.2 percent faster task completion and 32 percent lower execution costs compared with the previous system (note.com).

These figures deserve an important caveat: they come from the company's own internal tests in one specific configuration, relayed via secondary coverage. No independent verification exists, and it is unclear how representative the figures are for typical agent tasks.

The platform also gains new triggers for automations: Manus 2.0 agents can be activated by email, calendar events, Slack messages and Notion updates (Thurrott.com). In practice, this means the agent no longer simply waits for the user to open an app, but can react to events in the tools the organization and user already use.

Cue: the agent as actor, not tool

The most striking element of the launch, however, is Cue. Each agent will, according to Butterfly Effect, have its own identity — email address, phone number, wallet and computer — so that it can send messages, make calls and pay within a budget set by the user, and complete tasks on its own machine (Thurrott.com).

That moves the agent from being a tool inside an interface to a kind of independent actor with its own channels. India Today highlights everyday scenarios, such as an agent that would be able to scan a restaurant's QR code and order on the user's behalf (India Today). The app is also designed for scenarios where multiple agents are added to a group chat and divide tasks among themselves, such as research, candidate drafting and document drafting (note.com).

Here the distinction between design and demonstration must be kept clear: payment, calling, ordering and multi-agent collaboration are the company's vision and design descriptions, not features that have been independently tested. There is so far no documentation of how reliable such agent tasks are in practice.

Background: independent again, with capital on the way

The launch comes during a complicated period for the company. According to Cryptobriefing, Chinese authorities in April 2026 blocked an acquisition attempt by Meta, reported to be worth roughly $2 billion. Manus resumed fully independent operations on September 1, 2026 — barely a month before the 2.0 launch (Cryptobriefing).

At the same time, Manus is reported to be attempting to raise $500 million at a target valuation of around $4 billion. This too has been described as rumor-based and unconfirmed by the company, and the valuation figures for the Meta attempt vary between Cryptobriefing's two articles. The company's revenue is said to have risen from around $90 million in annual run-rate revenue in August 2025 to around $125 million in December 2025 — an increase of roughly 39 percent over four months, according to the same source, but with no named primary source or financial statements (Cryptobriefing).

Among the investors named are Benchmark, Tencent, HSG and ZhenFund, with Xiao Hong as chief executive and Ji Yichao as chief scientist. The company was founded in China and is now based in Singapore under the name Butterfly Effect Pte. Ltd. (Cryptobriefing; India Today).

What remains to be answered

The biggest open question concerns performance first: the Cascade figures of 23.2, 28.2 and 32 percent are internal test numbers, and it remains to be seen whether the improvements hold up in independent comparisons and in real work tasks. Then comes the question of reliability: an agent that calls, pays and orders on the user's behalf places entirely different demands on fault tolerance than one that writes documents, and none of these capabilities has been verified so far.

Add to that the commercial uncertainties. The ~$4 billion valuation, the funding round and the terms surrounding the blocked Meta acquisition all rest on secondary coverage flagged as "reportedly," and the figures do not fully align between sources. And the launch date itself — September 28 or 29 — is, for now, a matter of time zones.

What is in place regardless: this is the first major product launch from Manus as an independent company, and it presents a concrete architecture change and a clear ambition that agents should have their own identity and room to act. How far that ambition is from everyday reality, no source can yet say with confidence.