Manus, video and remote tasks: what the sources confirm—and what remains unproven

What the available sources actually say

This news proposal aims to explain features attributed to Manus 2.0 and Cue, including video editing, remote tasks and an app for creating assistants. However, the supplied sources include no announcement of either product, release notes or documentation describing Cue. Nor do they include a verifiable demonstration of the features attributed to Manus 2.0. On the basis of this material, those items cannot be presented as confirmed new features.

What the sources do provide is general product information. Manus’s Google Play listing describes an agent that breaks a task into subtasks, carries them out and delivers a result. Manus’s tools page says users can describe a task for the agent to execute, and states that it can generate different types of files. These are descriptions from the provider or app listing, not independent proof that every task can be completed from start to finish.

The official website also has a tool for creating videos from text. This confirms that there is a product page devoted to that capability. By itself, it does not show that the system can edit existing videos, that the capability is part of Manus 2.0, that it is available to all users or that Cue includes it. Generating a video from a prompt and modifying audiovisual material supplied by a user are different tasks; the source reviewed does not establish that they are equivalent.

Generating video is not the same as editing it

The distinction matters because phrases such as “video features” can cover very different capabilities. A text-to-video tool may produce a clip from a written prompt. Editing, by contrast, could involve working with an existing file, selecting clips, rearranging scenes, adding elements or exporting a modified version. The page reviewed supports only the first, general description: creating videos from text.

No input-and-output examples, format specifications, available controls or terms of use are provided. The sources also do not explain whether a person can review each step, correct the result while the process is underway or restore an earlier version. Without those details, saying that the agent “edits videos” or can handle a complete production would go beyond the available evidence.

The same caution applies to examples of tasks. The Google Play listing and tools page present the agent as capable of organizing and carrying out work, but they do not identify specific tasks whose completion has been verified by third parties. A description of how a system is intended to work is not a reproducible test, nor does it establish a success rate.

Remote tasks: access and permissions remain unclear

The idea of an agent working remotely raises practical questions that a general description of automation does not answer. To understand what it can do, readers would need to know whether it operates in its own environment or within external services; what authorizations it requests; whether it can read, change or send data; and how actions with consequences for the user are approved. The available pages and app listing do not explain these points.

Manus’s homepage presents it as a system for carrying out tasks and automating workflows. That describes the service’s positioning, but it does not specify the scope of its access or demonstrate that the agent can control external accounts. The app listing refers to a “computer” of its own for breaking down tasks, but the supplied information does not explain its configuration, isolation, connectivity or permissions either.

As a result, these sources cannot establish that Manus 2.0 or Cue can act remotely within third-party services, or what oversight controls they provide. They also do not clarify whether particular actions require explicit confirmation or whether users can consult an activity history. Before using an agent with sensitive information, these conditions should be checked in current product documentation and in the settings shown during authorization.

Availability and autonomy: what the evidence cannot establish

The presence of a Manus listing on Google Play establishes that an app page exists in that store; it is not enough to determine which features are enabled for each account, country or version. The sources do not say whether the purported Manus 2.0 or Cue features are in early access, rolling out gradually or generally available. They also provide no dates, subscription requirements or availability matrix.

Assessing autonomy takes more than checking whether an agent produces a response or file. To determine whether it completes a task from start to finish, tests would need to specify the goal, permitted tools, input data, actions taken and criteria used to judge the result. They should also record human interventions, errors, cases in which the agent stopped, and whether it can recover without making the situation worse.

None of the supplied sources is an independent performance evaluation. They therefore cannot establish reliability, compare Manus with other agents or confirm that it can operate without supervision. That does not prove such capabilities do not exist; it means the material reviewed does not substantiate them. The distinction matters: a lack of evidence in these sources is not proof of an inability.

What is missing to confirm Manus 2.0 and Cue

To turn the claims in the proposal into a confirmed news report, primary sources would need to identify Manus 2.0 and Cue explicitly: an official announcement, feature documentation or a demonstration attributable to the provider. For video editing, the evidence would need to show changes being made to existing material, rather than only generating a video from a prompt. For remote tasks, it should clarify the execution environment, supported services and required permissions.

It would also be necessary to distinguish announced availability from actual availability. A feature described on a page may be limited by country, account, version or rollout stage. The information reviewed here cannot resolve that. Likewise, a controlled demonstration can illustrate a workflow, but does not establish that it works reliably across a range of cases.

For now, the responsible conclusion is limited: Manus is described as an agent that carries out tasks and generates files; an official tool for creating videos from text exists; and the app listing describes a workflow that breaks tasks into subtasks. The supplied sources do not support attributing those capabilities specifically to Manus 2.0 or Cue, or claiming complete autonomy, general remote access or universal availability. Evaluation should remain open until verifiable documentation and independent testing are available.

Keep exploring

Corrections and transparency

If you spot incorrect or outdated information, send us a correction with the page and source we should review.