July 2026

The agent chat in the Interface tab now takes attachments. Click the paperclip, paste an image from your clipboard, or drop files onto the composer. Images become thumbnails you can hover to enlarge, other files become chips, and both stay on the message when you reload the page.

• Attached images now reach models that can see them. Images used to be dropped on the built-in LLM agent, so a model that could read them never got them. The five CLI coding agents fetch attachments from their sandbox instead.
• Attaching an image to an image or video model uses it as input rather than describing it: image models treat it as a reference image, and Veo and Kling use it as the first frame of the video when you have not set one.
• You can send an attachment with no text at all.

A WhatsApp QR connection can die on the provider's side, when the phone unlinks the device or WhatsApp logs the session out. Incoming messages stop arriving and nothing used to say so. NoClick now emails you when a connection drops, and every surface that reads your credentials reports it.

• Re-scanning the QR code reconnects the credential you already have instead of creating a second one, so you are not charged the $0.99 monthly connection fee twice. Scanning the same phone again also rebinds to your existing credential. Repeated fresh scans were what got some accounts logged out of every linked device.
• A credential whose connection has dropped is labeled Disconnected in the credential picker, and the node's Credentials tab tells you to re-scan that same credential.
• The new Channel disconnections toggle in Settings → Notifications controls these emails.
• Running a WhatsApp Receive Message step by hand on a dead connection now fails with re-scan instructions. On a healthy connection it reports that there is no live event instead of returning empty data that looked like a successful delivery.
• Ask an AI agent or the AI builder about a workflow and it now sees which of your connected accounts are disconnected, instead of reporting them all as fine.

Pressing Run now opens a popup that lists every step still missing something: a required field, an action, a connected account, or the operations an agent is allowed to call on a tool. You fill each one in the popup, using the same controls the config panel uses, and connecting an account no longer closes it and costs you your place. Nothing is blocked: the button reads Run anyway while anything is unfilled, and turns into Run once it is all there. The same check runs before a single node run and before Run from here, covering the nodes those runs actually touch.

When a workflow has more than one entry point, the popup lists them so you can untick the branches you do not want this time. Each row names what it leads to, such as "Then Slack, Gmail" or "Uses Telegram". An agent entry point comes with a box for its opening message, prefilled from the node's own message, and when that agent is the only thing you run, its answers land in the chat, which opens as soon as the run starts.

When the AI builder asks you to pick a resource like a board, a database, a repository, or a warehouse, and there is a way to create a new one, the question now offers a Create new shortcut. Click it and the builder makes the resource for you and fills it in, so you do not have to leave the chat to set one up first. This used to work only for Google Sheets and now covers resource pickers across the node catalog.

Agents can now propose a change to the workflow they run in. The suggestion arrives in the chat as a Proposed workflow edit card. Choose Send to builder to open the builder with the change ready to apply, or Dismiss to skip it.

When you drag a node from the palette, you can now drop it straight onto a target to wire it up. Drop it on an agent to add it as a tool, on the plus on an edge to splice it into that connection, or on a node's trailing plus to chain it after that node. Dropping on empty canvas still places the node on its own.