Skip to main content

Inbound Requests

Handling workflow requests from external sources

Workflow requests may be received by a thunk in different ways:

  • Each thunk has an Email Inbox so that email messages can be sent to it along with attachments. Each of these messages can be processed by custom AI Instructions to extract information and create one or more workflow requests. See Send requests to your thunk by email.

  • Each thunk also has a secure Webhook endpoint which can be used to route form submissions to, or send events from external applications. Each webhook also comes with a form link that people can open in a browser to send in a request without a Thunk.AI account. See Receive requests through a webhook or a shared form.

  • For deeper programmatic integrations, external applications can use a REST API to submit workflow requests, check on their progress, and retrieve their results.

  • For chat, a workflow with a single conversation input gets a ready-made hosted Chat UI that end users open from a link. To build chat into your own website or app instead, use the conversational interface via a REST API. Both run on the same conversation.

  • Every thunk also has a chat-based AI agent that can act upon natural language instructions to create new workflow requests (perhaps by loading them from a spreadsheet or by searching the web, etc.)

  • Finally, every thunk also provides interactive user interfaces to load individual workflow requests via a form, or to bulk-load them from files or folders. Adding work items with AI, with a form, or by bulk load are all choices on the Add menu of the work items list — see How to work with your work items list.

Separately, while debugging, you can probe a model the thunk already uses from a script — that path is not for production.

Find a message

Open Inbound Requests and pick a channel's tab. The tab lists every message that arrived on that channel, newest first, with its Status (Routed, Failed, In progress or Not routed), when it was Received, its Subject, who it is From, and an Excerpt.

Use the table's filter to narrow the list — for example, to messages that Failed, messages received after a date, or messages whose subject or sender contains some text. The filter is part of the page's link, so you can share a filtered view. Click a message to open its page.

If you can administer the thunk, the channel's settings are in the bar at the bottom of the tab: Instructions for every channel, plus Security for email, Test for a webhook, and API Keys for the API channel.

View a message

Each message a thunk receives has its own page. Open it from its channel's list, or from a work item's Created from link, to see exactly what arrived and what the thunk did with it. The top bar shows where the message came from — Inbound Requests › (channel) › (subject) — and clicking the channel returns to that channel's messages.

The page shows:

  • Input payload — the message as it arrived: an email as the email, a webhook's JSON body laid out for reading, a PDF in the viewer. Download content saves the original.

  • Details — the channel, the type of message, the sender, when it was received, and its message ID.

  • Routing — the work items the message was routed to and why, or the reason it was not routed. Expand Routing agent activity to follow the AI's steps.

Use Route Message to have the AI route the message again or to route it to a work item yourself, and Delete to remove it.


#

Did this answer your question?