Manus

What it does

Writes code

YesGenerates frontend, backend, database and auth for a web app from a conversation.Manus generates production-ready code including frontend, backend, database, and user authentication—all through simple conversation.manus.im/features/webapp ↗

Debugs and fixes

YesRuns what it builds in its own browser, finds the failure and fixes it without being asked.With its integrated browser, Manus can actively test what it builds — launching the application, interacting with it like a real user, detecting…manus.im/blog/manus-1.5-release ↗

Writes tests

PartlyClaims it tests the app it just generated; no claim that it authors test code or a test suite.Generate a full-stack web app with user login, database, and backend, then build, launch, and test it automatically.manus.im/download ↗

Whole-codebase work

NoVendor positions repository-scale work as the competitor's job; Manus delivers a whole app, not edits across your codebase.Write, edit, and manage the codebase yourselfmanus.im/compare/manus-vs-replit ↗

Ships it

YesDeployment is part of the deliverable, including backend infrastructure and database.Manus delivers the complete application as a single output. Frontend, backend, database, and deployment—all handled together, not pieced apart.manus.im/compare/manus-vs-replit ↗

Not established: Reviews code

How much it does unattended

Runs autonomously

YesTasks keep running in the cloud after you disconnect, including scheduled and background runs; the API's default mode does not stop to ask.execute tasks asynchronously in the cloud, including Scheduled Tasks and background operations that continue running even when you close your browser…manus.im/privacy ↗

Multi-agent

YesWide Research decomposes a task across many parallel sub-agents and a main agent synthesises the results.Instead of using a single AI agent that processes items sequentially, Wide Research deploys hundreds of independent agents that work in parallel.manus.im/docs/features/wide-research ↗

Agent permissions

PartlyAuthorisation is per task and per connector, with watch-and-interrupt for the browser and confirmation gates on terminal commands and high-credit runs, but not per-action approval for cloud tasks.One-time authorization: Each task requires your explicit permission before Manus can control the browsertrust.manus.im/faq ↗

Which models it runs on

Claude

PartlyClaude is selectable only for LLM features inside apps Manus builds for you; you cannot choose it for the Manus agent itself, and Anthropic is not among the disclosed foundational-model subprocessors.LLM-powered features can run on any of ten models across three providers: OpenAI (GPT-5), Anthropic (Claude), and Google (Gemini).manus.im/docs/website-builder/ai-capabi… ↗

GPT

PartlyOpenAI is a disclosed foundational-model subprocessor for Manus itself, and GPT is nameable for LLM features in generated apps, but the agent's model is never user-selectable.OpenAI United States Foundational modelstrust.manus.im/subprocessors ↗

Open models

NoThe disclosed subprocessor list names only Google Cloud Platform and OpenAI as foundational-model providers; no open-weight model is named anywhere as running Manus.Google Cloud Platform United States Cloud infrastructure and foundational modelstrust.manus.im/subprocessors ↗

Model choice

NoThe vendor picks the model; the API's agent_profile chooses a capability tier (lite/standard/max), not a model.Being LLM-agnostic, Manus selects the right model for each use case so users don't have to.trust.manus.im/faq ↗

Where you use it

In your editor

NoNo editor or IDE extension; the vendor's own surface list is desktop and mobile, plus a Chrome/Edge browser extension.The AI agent that doesn't just think, it delivers. Available for macOS, Windows, iOS and Android.manus.im/download ↗

On the command line

NoNo terminal client is offered; the download page enumerates desktop and mobile apps only.The AI agent that doesn't just think, it delivers. Available for macOS, Windows, iOS and Android.manus.im/download ↗

In your pipeline

NoGitHub appears only as an OAuth data connector; no CI, Actions or pull-request surface is claimed, and the vendor's stated deliverable is a deployed app rather than a diff.Available Connectors: Google Drive, Gmail, Google Calendar, Notion, Slack, GitHub, Stripe, HubSpot, Hugging Face, and more.trust.manus.im/faq ↗

In a browser

YesThe primary surface is the hosted web app at manus.im, where connectors and settings are managed.You can revoke access at any time from the integrations page on manus.im.open.manus.ai/docs/v2/connectors ↗

Whose machine it runs on

Self-hosted

NoTasks execute in Manus-operated cloud sandboxes on GCP, Azure and AWS; no on-premise or self-hosted deployment is offered anywhere on the site.execute tasks asynchronously in the cloud, including Scheduled Tasks and background operations that continue running even when you close your browser…manus.im/privacy ↗

Not established: Open source

What it costs to run

How it meters

YesMetered in credits consumed by LLM tokens, sandbox VMs and third-party APIs; Team adds a per-seat price on top.Credits are primarily consumed based on: LLM tokens... Virtual machines: Used for cloud environments that support file operations, browser…help.manus.im/en/articles/11711097-what… ↗

Free tier

YesPermanent $0 plan with 300 daily refresh credits, 1 concurrent task and 2 scheduled tasks; no call required.[Free Plan Price: $0/month] 1. Access to Chat Mode and Manus 1.6 Lite in Agent Mode.help.manus.im/en/articles/11711111-what… ↗

API access

YesSelf-serve REST API keys from account settings, up to 50 per account; free accounts are downgraded to the lite profile and OAuth "Open Apps" need a Team account.Click Create API Key and give it a descriptive name (e.g. "production", "dev-testing"). Each account can have up to 50 API keys.open.manus.ai/docs/v2/authentication ↗

MCP server

PartlyManus is an MCP client, with prebuilt and custom MCP servers on all plans, but publishes no MCP endpoint for another agent to call it; the inbound path is the REST API.When you connect an MCP server, Manus can read structured data (like text from docs, issues, or diagrams) and, where supported, perform limited…manus.im/docs/integrations/mcp-connecto… ↗

Bring your own key

NoNo bring-your-own inference: models are not selectable, and built-in data sources explicitly need no keys of your own.All data sources are included with your Manus subscription. No additional API keys or setup required.manus.im/docs/integrations/data-sources ↗

Buying it for a team

A company can buy it

YesTeam is a multi-seat plan sold to organisations, with unified billing and invoices emailed to the purchaser.designed for organizations that want to use Manus across their workforce, allowing multiple team members to independently access and utilize our…help.manus.im/en/articles/11711750-what… ↗

Seat model

YesFrom $20 per seat/month; guest users are unlimited and consume no seat; no minimum or maximum seat count is published.[Team Plan Price: Starting from $20 per seat/month (Enjoy a 17% discount with annual billing)]help.manus.im/en/articles/11711111-what… ↗

Pooled budget

YesEvery billed seat contributes to one shared team credit pool; heavy users draw on lighter users' unused credits.Each billed user adds credits to your team's shared pool. This eliminates waste and maximizes your subscription value.manus.im/team ↗

Admin controls

PartlyTeam-only: four roles, a Security policies panel governing how members may join and access, and admin control over data migration. There is no policy engine constraining what the agent may do.a "Security policies" section will appear for Team Owners and Super Admins to control how members access Manus Team after SSO is enabled.help.manus.im/en/articles/12807937-wher… ↗

Audit log

PartlyAudit logs are claimed as a Team feature; the help center documents only a timestamped per-user credit-consumption activity log.Audit logs and enterprise-grade compliance (SOC 2 Type II, ISO 27001)manus.im/team ↗

Single sign-on

PartlyTeam plan only, configured through WorkOS against your identity provider, as a paid $150 add-on waived at 30+ seats; the vendor never names SAML or OIDC.Below 30 seats, your SSO subscription will be charged a flat fee of 150 USD + tax.help.manus.im/en/articles/12697595-what… ↗

What happens to your code

Opt out of training

PartlyA named opt-out control is a Team plan feature; separately the vendor claims platform-wide that model providers are contractually barred from training on your data.Data Training Opt-Out: Supports opting out of data training to protect privacy.help.manus.im/en/articles/11711111-what… ↗

Data residency

PartlyStored in the US and Singapore with SCCs for transfers; no customer-selectable region and no EU hosting option.Data storage location: - US - Singapore For international transfers, we use EU Standard Contractual Clauses (SCCs) and other appropriate safeguards…trust.manus.im/faq ↗

Getting out

PartlySelf-serve backup export of account and task data and self-serve account deletion, but monthly credits are cleared every cycle and reset to zero when a subscription ends; add-on credits survive only while a paid subscription is active.Manus does not allow unused credits to roll over to the next cycle. At the start of each new cycle, all unused credits are cleared and replaced with…help.manus.im/en/articles/11711097-what… ↗

Certifications

YesISO/IEC 27001:2022, ISO/IEC 27701:2019 and SOC 2 Type 2 for Security; GDPR is addressed by controller/processor roles, a DPA and EU/UK representatives rather than a certification, and HIPAA data is prohibited input.Manus holds ISO/IEC 27001:2022 (Information Security Management System) and ISO/IEC 27701:2019 (Privacy Information Management System)…trust.manus.im/faq ↗