Skip to content

Describe it. Watch it build.

GoRunner AI turns a sentence into a running flow: it picks the pieces, wires the steps, maps the data, and deploys. Then it stays around to edit by chat.

GoRunner AI⌘K

Trigger: Gmail, new email in the support inbox
AI step: classify intent and draft a reply
Action: Salesforce, create a case with priority
Action: Slack, alert #support with the draft
GmailGoRunner AISalesforceSlack
 livefirst run 412ms

Three real scenarios, on loop.

Builds whole flows

From "when a support email arrives, classify it and open a case" to a deployed flow with the right trigger, mapped fields, and a sensible branch for edge cases.

Edits by chat

"Add an approval step before the refund." "Swap Gmail for Outlook." "Why did run 9f2jx fail?" The copilot edits and explains the flow you already have.

Works inside flows too

AI steps classify, extract, draft, summarize, and decide as part of a run, using GoRunner AI or your own keys for Claude, OpenAI, Gemini, and Mistral.

Powerful, with the brakes on

An AI that can touch your CRM and your customers needs more than capability. It needs manners.

Drafts, never production

Every AI edit lands as a draft you can diff and reject. Nothing the model does touches a live flow until a person deploys it.

Approvals where it matters

Put a human-approval step in front of anything irreversible: refunds, deletions, customer-facing sends. The model proposes; people dispose.

Every call on the record

AI steps log their inputs, outputs, and confidence like any other step, so audits read the same whether a person or a model did the work.

an AI step inside a flow

Invoice attachedExtract line itemsReview if confidence < 0.9Create invoice

Say it. Ship it.

Your first described flow is live in the time it took to read this page.

Free plan · 1,000 runs a month · self-host or cloud