Create and drive Manus AI agent tasks, poll their progress, manage projects, files, webhooks and published websites, over Manus's API v2.
Manus ships in the w6w first-party pack. It declares 32 actions, 2 health checks, and the host runs its code in a sandbox that never sees the credential.
io.w6w.manusManus is a general-purpose AI agent platform — it browses, writes code, builds and publishes websites, and works through multi-step tasks the way a person would, rather than answering in a single turn. This app drives Manus’s API v2: create a task from a prompt, poll its event stream for progress, send follow-up messages, and confirm the actions a task pauses on when it needs a human’s go-ahead.
Beyond the core task lifecycle, it covers the platform features a workflow actually touches: projects (shared instructions applied to every task inside them), file uploads a task can read, webhooks as an alternative to polling, the connectors and skills a task can be configured with, custom agents’ subtasks, a team’s online browser clients (for the human-in-the-loop browser-handoff flow), and the credit/usage accounting that decides whether a task can run at all. It also covers the websites a Manus task can build: checking publish status, browsing checkpoints, and publishing or updating one.
Manus’s docs describe two API generations. v1 is explicitly deprecated in the vendor’s own words (“API v1 has been deprecated and will be removed in the future”) — every action here is built against v2 only.
Three routes to the same 32 actions. The Workflow tab is generated from Manus's own manifest and carries its real ids, so it is copy-pasteable; the Code and CLI examples are the same call for any action on any app, so every app-specific value in them is a blank you fill in.
browser-online-list List the account's currently-online browser clients.
task-list-messages List a task's event history: messages, status changes and (verbose) agent activity.
usage-available-credits Get the account's (or team's) spendable credit balance and refresh schedule.
usage-team-log List each team member's task count and credit consumption over a date range.
usage-team-statistic Get the team's daily credit consumption over a date range.
webhook-public-key Get the public key used to verify webhook notification signatures.
A workflow step names the app and the action, and the editor fills in the
connection when you pick one. This is the Step shape from the
workflow spec, carrying Manus's real ids.
{
"manifestVersion": "2",
"name": "manus-example",
"steps": [
{
"id": "agent-detail",
"uses": {
"app": "io.w6w.manus",
"action": "agent-detail",
"connection": "conn_YOUR_CONNECTION_ID"
},
"with": {
"agentId": "<agentId>"
}
}
]
}agent-detail agent-list agent-update browser-online-list connector-list +27 more actions available
Every app-specific value here is a blank you have to fill in. An
app action is reached through the connection that authenticates it, so the
address is a connection id, not the app id — and connections belong to your account,
so a public page cannot know yours. Create one for Manus, then fill in
the three blanks: conn_YOUR_CONNECTION_ID, the action key, and the
parameters that action declares. The call itself is real — the shape is transcribed
from the studio's own snippet builder, which prints the same kind of blanks — but
nothing in it is specific to Manus. The Workflow tab is where this app's
real ids are.
npm install @w6w/sdkyarn add @w6w/sdkpnpm add @w6w/sdkdeno add npm:@w6w/sdkimport { W6wClient, isActionRun } from "@w6w/sdk";
// Reads W6W_BASE_URL and W6W_TOKEN from the environment when omitted.
const client = new W6wClient();
const envelope = await client.run({
urn: "conn_YOUR_CONNECTION_ID",
action: "agent-detail",
payload: {
agentId: "<value>",
},
});
if (isActionRun(envelope)) console.log(envelope.value); npm install -g @w6w/cli w6w run conn_YOUR_CONNECTION_ID --action agent-detail --payload '{"agentId":"<value>"}' Give an AI agent Manus — without giving it Manus's credentials. One MCP endpoint exposes every app, function and workflow the caller is entitled to, as tools it can discover and run. Access is granted per team while we onboard.
One tool call{
"name": "w6w_invoke",
"arguments": {
"ref": "app:io.w6w.manus#agent-detail",
"input": {
"agentId": "<agentId>"
}
}
}
Every tool names its target with a single ref. The
app: form above doesn't name a connection at all — the
host resolves which of the caller's Manus connections to sign
with, and refuses rather than guesses when the answer is ambiguous.
The token is attached host-side, at the moment of the call. It is never a tool argument, never in the model's context, and never in a transcript — so a prompt injection has nothing to exfiltrate.
Tools are derived per end user from what that person has actually connected and is entitled to — not one shared bot identity carrying the union of everyone's access.
Multi-step work runs on the workflow engine and returns a run handle the agent can poll — retries, branching and state survive the conversation that started them.
Manus's declared health checks are on the surface too, so an agent can tell "the vendor is down" from "your credential expired" before it burns a retry on either.
The MCP surface is part of the hosted platform. Manus itself is MIT, and the runtime that executes it is source-available (FSL).
Manus declares its own checks, so its health is a property of the app rather than something the host guesses at.
Component status from status.manus.im, scoped to the API and the task-execution ('manus computer') components — not the web app this app never calls.
Spendable credit balance, read from GET /v2/usage.availableCredits.