August 2026
Everything that changed in NoClick in August 2026.
Improvements and fixes1
The Next Run countdown on a schedule keeps counting down after a run fires. On a schedule that fires faster than the countdown could refresh itself, it could stop at Running... and stay there until you reopened the node.
Schedules follow your timezone through daylight saving
Run times, time windows, and day boundaries are now evaluated directly in the timezone you pick on the node. A 9:00 AM schedule fires at 9:00 AM local time year round instead of moving by an hour when the clocks change.
Improvements and fixes2
Restoring a workflow from the trash or from a version checkpoint brings back every schedule on a Schedule node, with its timezone and its limits intact. A node holding several schedules used to come back with one.
Time windows are inclusive at both ends, so a schedule that runs to
6:00 PMstill fires the 6:00 PM tick.
Improvements and fixes1
The Copy label in the Setup tab's connect step no longer paints over long terminal commands. The command scrolls underneath the label instead, and its last characters stay readable.
Improvements and fixes4
Files in the File Upload block are listed as one contained list, each row with its size and a Copy link button that copies the file's permanent URL. The last row's remove button is clickable again, having sat underneath the block's resize handle.
Template pages for a specific coding agent, such as a template with Codex, load again instead of returning a not found error.
Permanently deleting a workflow stops any run still in progress first. If a run does not stop in time the workflow stays in the trash and you are asked to try again, so a deleted workflow cannot keep acting.
A channel disconnection is emailed once. Two copies could arrive when the provider delivered the same event twice.
Improvements and fixes3
Template pages, shared workflow links, and workspace invite links load again.
Interfaces the AI builder writes come out filled in again. Since August 20 the builder had been finishing the shape of a workflow without filling in the nodes it added, which left interface components empty.
The Test Run screen no longer spins forever when an agent replies with structured output, and it starts the test run you asked for instead of the first one on the workflow.
Improvements and fixes1
Keyboard shortcuts work again in the workflow editor. A few seconds after a page load they stopped responding, so
N, theWILSVview switches, arrow key selection,Enter,Esc,Shift+Rto run the workflow, the config panel'sXandY, andCmd+Enterto run a node all did nothing.[in the workflow list was affected too.
Improvements and fixes1
The AI builder no longer reports a build as finished unless it confirms completion itself. A run that stops early or reaches its turn limit now ends as incomplete and keeps the changes it already made, instead of reporting success on a canvas it never touched.
Improvements and fixes1
Errors now say what actually went wrong. When a page failed, the error screen reported a fault in itself rather than the underlying problem, which also left connecting an MCP client while signed out showing "Unexpected Server Error" instead of taking you to the sign-in page.
Improvements and fixes1
A part of the app that fails to load in the background just after a release now recovers with a single reload rather than leaving the page broken. Safari and Firefox describe that failure in wording NoClick did not recognize before. The reload happens at most once every 30 seconds, so a load blocked for some other reason cannot loop.
Improvements and fixes3
A node that could not be deleted deletes again. Selecting it and pressing
Backspace, or using its delete button, did nothing, and undo and redo stopped responding on that canvas until you reloaded the page. A node that reached the canvas without a position broke both, and one is now repaired as it arrives.An agent the AI builder runs to test it keeps the tools wired into it. The run dropped the connection between the agent and the services feeding it, so the agent executed with none of the actions you had allowed and reported that it had no tools.
The check that stops the builder from ending a turn after only describing what it is about to do now applies at any point in a build. It stood down once it had fired earlier in the conversation, and read a message sent mid-build as a question rather than as work still owed.
Scheduled triggers catch what arrives while you set them up
Triggers that poll on a schedule, like Google Forms On Form Response, Google Sheets On New Row, and Gmail Poll for New Emails, now start watching as soon as their configuration is complete and treat anything that arrives after that as new. Before, the schedule could be created before the trigger had its required fields, which produced a failed run every interval while you were still filling in the panel, and the first poll after that counted everything already there, including the test entry you had just submitted, as history.
- The schedule is created only once the trigger's required fields are filled in, so ticks during setup no longer appear as failed runs in the Logs tab.
- A turn where the AI builder or the MCP server fills in the last required field now registers the trigger, so "your trigger is set up" means it is listening.
- Pointing a trigger at a different form, sheet, or mailbox starts a fresh baseline for the new resource instead of firing for everything already in it. Edits that do not change what the trigger watches, including the ones the AI builder makes while finishing a workflow, no longer reset what it has already seen.
- Running the workflow yourself delivers newly polled items to an AI agent wired to the trigger, the same as a scheduled run. A poll that finds nothing new still runs nothing downstream, manual runs included.
- Changing a node away from its trigger operation removes its schedule, and a trigger whose registration never landed is repaired on its own within a day.
Improvements and fixes1
Microsoft credentials are named for the service you connected, so Word, Excel, OneDrive, and the rest each count separately against your per-service credential limit instead of sharing one.
Improvements and fixes3
The AI builder no longer ends a build after only describing what it is about to do. A first message that produced a plan, no nodes, and no error now continues into the work. The check also applies right after you answer one of the builder's questions, which is where a run was most likely to stop with your answer unused.
Asking the builder for an agent on a specific model now sets that model. Naming Claude used to give you an agent called "Claude" running the default model instead. Agents are also no longer named after a model vendor, since that name goes stale as soon as you switch models.
Credentials you connect while the builder is working stay attached. A connected account could disappear from the workflow seconds later, and the builder would then ask you to connect it again.