Aidbase AI support: manage knowledge sources, chatbots, email inboxes and ticket forms, get chatbot replies, read chats, emails and tickets, and reply to, update or block them.
Aidbase ships in the w6w first-party pack. It declares 43 actions, 3 health checks, and the host runs its code in a sandbox that never sees the credential.
io.w6w.aidbaseAidbase builds AI support agents trained on your own content. This app drives its REST API from a workflow: create and train knowledge sources (websites, YouTube videos, uploaded documents and FAQs), attach them to chatbots, email inboxes and ticket forms, and keep the answers current without opening the dashboard.
On the conversation side it asks a chatbot for an AI reply (continuing a session), lists and reads chats, inbox emails and form tickets over a date range, and acts on them: reply to an email or ticket, change a status or priority, or block an abusive sender.
Authentication is a single Aidbase API key sent as a bearer token. Keys carry scopes chosen when the key is created, so what a workflow may change follows the key’s permissions.
Three routes to the same 43 actions. The Workflow tab is generated from Aidbase'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.
account-status Check that the API key is accepted. Returns the API status; the masked key Aidbase echoes back is dropped.
chatbot-chat-block Block a chat so its user can no longer interact. Needs an API key with the CHATBOTS_WRITE scope.
chatbot-chat-list List the conversations a chatbot has had, optionally within a date range. Paginated.
chatbot-knowledge-list List the knowledge items attached to a chatbot. Paginated.
chatbot-knowledge-remove Detach a knowledge item from a chatbot. The knowledge item itself is not deleted.
chatbot-reply Send a message to a chatbot and get its AI reply. Pass the `session_id` from a previous reply to continue that conversation.
email-block Block an email so its sender can no longer interact. Needs an API key with the EMAILINBOXES_WRITE scope.
email-inbox-knowledge-add Attach a trained knowledge item to a email inbox.
email-inbox-knowledge-list List the knowledge items attached to a email inbox. Paginated.
email-inbox-knowledge-remove Detach a knowledge item from a email inbox. The knowledge item itself is not deleted.
email-list List the emails in an inbox, optionally filtered by status and date range. Paginated.
email-reply Send a reply to an email on behalf of the inbox owner. Needs the EMAILS_WRITE scope.
email-update Change an email's status and/or priority. At least one is required, and Aidbase rejects any other field. Needs the EMAILS_WRITE scope.
knowledge-delete Delete a knowledge item. Chatbots, inboxes and forms that used it lose it.
knowledge-document-create Start a document knowledge item. Returns a signed `upload_url`: upload the file there, then run Finalize Document.
knowledge-document-finalize Complete a document knowledge item after its file has been uploaded to the `upload_url`.
knowledge-faq-item-create Add a question and answer to an FAQ knowledge item. Categories that do not exist are created. Run Train Knowledge to index the change.
knowledge-faq-item-delete Delete one question/answer item from an FAQ knowledge item.
knowledge-faq-item-list List the question/answer items of an FAQ knowledge item. Only applies to items of type faq.
knowledge-list List the knowledge items (websites, videos, documents, FAQs) in the account. Paginated.
knowledge-sub-page-list List the sub-pages crawled for a website knowledge item. Only applies to items of type website.
knowledge-train Start a training run for a knowledge item. It finishes later: poll Get Knowledge and watch `is_training` and `trained_at`.
knowledge-website-create Create a website knowledge item. Run Train Knowledge afterwards to crawl and index it.
ticket-block Block a ticket so its submitter can no longer interact. Needs an API key with the TICKETFORMS_WRITE scope.
ticket-form-knowledge-add Attach a trained knowledge item to a ticket form.
ticket-form-knowledge-list List the knowledge items attached to a ticket form. Paginated.
ticket-form-knowledge-remove Detach a knowledge item from a ticket form. The knowledge item itself is not deleted.
ticket-list List the tickets submitted to a form, optionally filtered by status and date range. Paginated.
ticket-reply Send a reply to a ticket on behalf of the form owner. Needs the TICKETS_WRITE scope.
ticket-update Change a ticket's status and/or priority. At least one is required, and Aidbase rejects any other field. Needs the TICKETS_WRITE scope.
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 Aidbase's real ids.
{
"manifestVersion": "2",
"name": "aidbase-example",
"steps": [
{
"id": "chatbot-chat-block",
"uses": {
"app": "io.w6w.aidbase",
"action": "chatbot-chat-block",
"connection": "conn_YOUR_CONNECTION_ID"
},
"with": {
"chatbotId": "<chatbotId>",
"chatId": "<chatId>"
}
}
]
}chatbot-chat-block account-status chatbot-chat-get chatbot-chat-list chatbot-get +38 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 Aidbase, 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 Aidbase. 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: "chatbot-chat-block",
payload: {
chatbotId: "<value>",
chatId: "<value>",
},
});
if (isActionRun(envelope)) console.log(envelope.value); npm install -g @w6w/cli w6w run conn_YOUR_CONNECTION_ID --action chatbot-chat-block --payload '{"chatbotId":"<value>","chatId":"<value>"}' Give an AI agent Aidbase — without giving it Aidbase'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.aidbase#chatbot-chat-block",
"input": {
"chatbotId": "<chatbotId>",
"chatId": "<chatId>"
}
}
}
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 Aidbase 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.
Aidbase'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. Aidbase itself is MIT, and the runtime that executes it is source-available (FSL).
Aidbase declares its own checks, so its health is a property of the app rather than something the host guesses at.
Unauthenticated GET https://api.aidbase.ai/v1/status. A 401 carrying Aidbase's JSON `{"success": false, "message": …}` passes — it proves the API and its auth layer are answering. Key validity is the derived `auth:*` check's job.