Landbot chatbot platform: read channels and customers, send messages and WhatsApp templates, assign conversations to agents or bots, manage customer fields and message webhooks.
Landbot ships in the w6w first-party pack. It declares 28 actions, 3 health checks, and the host runs its code in a sandbox that never sees the credential.
io.w6w.landbotLandbot is a no-code conversational platform: bots that talk to customers over web chat, WhatsApp and Messenger, with human agents able to take a conversation over. This app drives its Platform API from a workflow.
It reads channels and the customers (conversations) in them, pulls a customer’s full message transcript, and sends text, images, locations and approved WhatsApp templates into a conversation. It also moves a conversation around: assign it to yourself, a specific agent or a bot (optionally starting the bot at a chosen node), unassign, archive, block or opt a customer out of WhatsApp.
Per-customer custom fields can be read, created, changed and deleted, so a workflow can feed the data a bot collected into the rest of a stack or seed what a bot should know. Message hooks register a URL that Landbot calls for every message on a channel.
Authentication is one workspace-wide agent token. Channel tokens and hook secrets that Landbot’s read endpoints return are never passed back out of this app.
Three routes to the same 28 actions. The Workflow tab is generated from Landbot'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.
channel-list List the channels (web, WhatsApp, Messenger, APIchat) in the workspace. Channel tokens are redacted.
customer-archive Archive a customer. Landbot refuses with 412 when the customer is assigned to another agent.
customer-assign Assign the customer to the agent that owns the token. Landbot refuses with 412 when another agent already has it.
customer-assign-bot Hand the customer to a bot. By default the bot starts at once at its start node; set a node to start elsewhere, or turn launch off to assign without starting.
customer-list List customers (conversations) in the workspace, filterable by channel, agent, archived and WhatsApp opt-in, with a name, email or phone search.
customer-messages-list Return the complete conversation transcript for a customer. There is no pagination and no guaranteed order: sort by message_datetime if you need it chronological. The set of message types is open-ended.
customer-opt-out Delete the customer's WhatsApp opt-ins. Landbot refuses with 412 when the customer has not opted in.
customer-unassign Remove the customer from its current agent. Landbot refuses with 412 when it is assigned to another agent.
message-hook-create Register a URL Landbot calls for every message on a channel. The optional token is sent back to your URL by Landbot; it is not returned by this action.
message-hook-list List the URLs Landbot calls for every message on a channel. Hook secrets are redacted.
send-template Send an approved WhatsApp template. Template ids come from List WhatsApp Templates; body parameters fill the {{1}}, {{2}} placeholders in order.
send-text Send a text message to a customer. A WhatsApp customer outside the 24-hour window needs a template instead.
whatsapp-template-list List the WhatsApp templates available to Send WhatsApp Template. With more than one channel, filter by channel to avoid duplicates.
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 Landbot's real ids.
{
"manifestVersion": "2",
"name": "landbot-example",
"steps": [
{
"id": "channel-get",
"uses": {
"app": "io.w6w.landbot",
"action": "channel-get",
"connection": "conn_YOUR_CONNECTION_ID"
},
"with": {
"channelId": "<channelId>"
}
}
]
}channel-get channel-list customer-assign customer-assign-agent customer-assign-bot +23 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 Landbot, 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 Landbot. 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: "channel-get",
payload: {
channelId: "<value>",
},
});
if (isActionRun(envelope)) console.log(envelope.value); npm install -g @w6w/cli w6w run conn_YOUR_CONNECTION_ID --action channel-get --payload '{"channelId":"<value>"}' Give an AI agent Landbot — without giving it Landbot'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.landbot#channel-get",
"input": {
"channelId": "<channelId>"
}
}
}
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 Landbot 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.
Landbot'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. Landbot itself is MIT, and the runtime that executes it is source-available (FSL).
Landbot declares its own checks, so its health is a property of the app rather than something the host guesses at.
Landbot's Statuspage: `Chats` decides down; the WhatsApp, Messenger, web-bot and Webhooks components can degrade the verdict but never take it down. Other components are detail.
Unauthenticated GET https://api.landbot.io/v1/channels/. A 401 carrying Landbot's JSON `{"detail": …}` passes — it proves the API and its auth layer are answering. Credential validity is the derived `auth:*` check's job.