Autonomous AI agent
How to use
Manus.
Delegating multi-step research, analysis, file creation, browser work, and small application or website builds.
Manus is a general autonomous AI agent. Delegating multi-step research, analysis, file creation, browser work, and small application or website builds. This guide covers the whole path in one place: official access, a first session that produces something reviewable, the checks that make output trustworthy, and the permissions worth limiting before you connect real work.
AI should make the work easier to inspect. If the workflow removes the source, the owner, or the review step, redesign the workflow.
Access & setup
Find, install, and sign in to Manus
Get into the official Manus experience with the right account and a setup you understand.
- 01
Start at https://manus.im/ and confirm the domain before entering account or payment information.
- 02
Availability: Official web application with task environments and connected messaging experiences on supported plans.
- 03
A supported browser and Manus account. Features, credits, persistent cloud environments, and team visibility depend on the active plan.
- 04
Sign in with the account you intend to keep using, then review plan, data, notification, and permission settings.
- 05
Run one low-risk test task before connecting sensitive files, repositories, or workspace data.
- Use official download pages.
- Review permissions during setup.
- Keep installers and applications updated.
First session
Your first useful Manus session
Learn the interaction loop using a small task with a clear outcome.
- 01
Open Manus and define one deliverable, its audience, evidence standard, and prohibited actions.
- 02
Provide only authorized context and ask to review the task plan before execution.
- 03
Monitor sources and tool actions, taking over whenever login, approval, or judgment is required.
- 04
Inspect every generated file and external change before accepting the completed task.
- State the outcome before the background.
- Provide the real source material.
- Review the result before expanding the task.
Quality control
Check the quality of Manus output
Establish that an autonomous run did the right thing, not merely that it finished.
- 01
Define what the run should achieve and what it must never touch before granting it a single tool.
- 02
Read the full execution trace: which tools were called, with what arguments, and in what order.
- 03
Verify the side effects directly in the target system rather than trusting the agent's own report of success.
- 04
Confirm failures surfaced as failures — a silent retry loop or a swallowed error is more dangerous than a crash.
- 05
Re-run the same task and compare: an agent that behaves differently across identical runs is not yet production-ready.
- Verify side effects in the system of record, not in the agent's summary.
- Require human approval for any irreversible or outward-facing action.
- Log every tool call so a run can be reconstructed afterwards.
Privacy & permissions
Use Manus safely
Bound what an autonomous system can reach before you let it run unattended.
- 01
Enumerate every tool, credential, and system the agent can reach, and remove the ones it does not need.
- 02
Require explicit human approval for irreversible actions: sending, publishing, paying, deleting, or deploying.
- 03
Run against non-production data until behaviour is predictable across repeated runs.
- 04
Set hard limits on spend, iterations, and runtime so a failure loop cannot run unbounded.
- 05
Treat anything the agent reads from the web or a document as data, never as instructions it may follow.
- Autonomous tasks can browse, create files, and operate in cloud environments. Understand task visibility, keep secrets out of prompts, constrain external actions, and verify all deliverables.
- Follow your organisation's approved-use policy.
- Never treat fluent output as authorization to act.
Core workflows
Step-by-step ways to use Manus for the work it does best.
Each workflow below has its own outcome, steps and review checkpoints.
Frame a multi-step assignment so Manus can research, analyze, and deliver a reviewable result.
↓ 02 WorkflowBuild and hand off a small website with ManusUse Manus's sandbox to create a focused site, then review its code, deployment, and ownership.
↓Workflow 01
Delegate a complex research task to Manus
Frame a multi-step assignment so Manus can research, analyze, and deliver a reviewable result.
Outcome: Receive a useful deliverable while retaining control over sources, browser actions, and conclusions.
- 01
Define the decision, deliverable, evidence standard, deadline, and actions Manus must not take.
- 02
Provide authorized inputs and ask Manus to show its plan before starting expensive or external steps.
- 03
Review the plan, narrow ambiguous branches, and require primary sources for consequential claims.
- 04
Monitor task progress and take over the browser when authentication, judgment, or approval is required.
- 05
Audit sources, calculations, files, and external changes before accepting or sharing the deliverable.
What good use looks like
- Set explicit stop conditions.
- Keep irreversible actions human-approved.
- Verify the final artifact, not only the summary.
Workflow 02
Build and hand off a small website with Manus
Use Manus's sandbox to create a focused site, then review its code, deployment, and ownership.
Outcome: Produce a maintainable website with clear control over files, credentials, and future updates.
- 01
Write the audience, pages, content, visual direction, functional requirements, and explicit exclusions.
- 02
Ask for the information architecture and technical plan before Manus writes or deploys code.
- 03
Build one complete flow with realistic content and test desktop, mobile, empty, and error states.
- 04
Inspect the generated files, dependencies, forms, analytics, secrets, accessibility, and legal content.
- 05
Export or connect the code, assign ownership, and use a controlled production deployment with rollback.
What good use looks like
- Own the source code.
- Never paste production secrets into a task.
- Test forms and failure paths manually.
Official references
Check the current product documentation.
Features, plan limits, availability, and data controls change. These official pages are the starting points used for this guide.