September 2026

Pick Direct Post Video or Direct Post Photo and the node settings now show a Post to TikTok panel: the connected creator's name, a preview of the media and caption, and the posting options TikTok currently allows on that account. Visibility is never filled in for you, comments, Duet and Stitch stay off until you turn them on, and an interaction the creator has turned off in TikTok cannot be enabled from here.

Every run re-checks those options against the account before it posts, so a visibility or an interaction the account no longer allows fails the run instead of publishing something more widely visible than you chose.

• Commercial content has its own controls. Disclose commercial content reveals Promote your own brand and Paid partnership with another brand, at least one of which has to be chosen, and a paid partnership cannot be posted privately. The workflow will not run while those settings contradict each other, and the panel links the TikTok policies you are agreeing to. AI-generated content marks a post as AI generated.
• The node output now carries published and publish_status alongside status, because TikTok accepts a post before it finishes processing it. If it is still processing, pass the returned publish_id to Check Video Publish Status rather than running Direct Post again, which posts a second time.
• Video you attach to the node is uploaded to TikTok directly, so it no longer needs a public address. Hosted video and photo URLs still have to be HTTPS on a domain verified for your TikTok developer app.
• Connecting another TikTok account now always shows TikTok's own consent screen. An already signed-in session used to be approved silently, which could reconnect the same creator when you meant to add a different one.

Pasting an API key used to save whatever you typed. A key with a stray line break, a web address pasted into a token field, or a key the service does not accept all saved looking connected, and you found out at the first run. NoClick now tidies the value and asks the provider about it before storing anything.

When the provider answers, the form shows what it said about your account rather than a tick: your own boards, sites, or workbooks, or Connected as and the account name. When the provider refuses, nothing is saved, and you get its own words along with what that usually means, such as an address serving a sign-in page instead of an API, or a key that reads as truncated or revoked. When NoClick cannot reach the provider at all, the credential saves and says Saved, not verified rather than claiming to be working.

• Problems are reported under the field they belong to. A web address pasted into a key field, placeholder text left in, and a value the service's own rules reject each name the field, and clear as you edit it.
• Credential forms now show the setup instructions and the per-field help for the service. Both were already written for each service and neither was displayed.
• A credential that proved itself when you connected it is remembered as Verified, with the account it connected as, so a node's Credentials tab says more than that something is attached.
• Tableau accepts your site address as it appears in your browser. Paste the whole URL into Server URL and NoClick fills in Site Content URL from it. The Cloud Manager console address is refused when you save, because it never serves the API, and the message names what to use instead.
• The same check runs wherever a key is entered: on a node, in Settings → Credentials, during Setup, and on a link someone sent you to connect an account for their workflow.

Moving an agent between Claude Code, Codex, OpenCode, OpenClaw, and Hermes used to end the thread it was on. Your next message started a new conversation with the recent turns pasted in as background. The conversation now moves into the harness you picked and carries on: the same thread, with each tool call still paired to what it returned, so the agent resumes where it left off rather than reading a summary of it.

• Any of the five CLI agents can hand a conversation to any of the others. The built-in LLM agent keeps its history in NoClick rather than in a session store of its own, so switching between it and a CLI agent still carries the recent turns across as context, about 4,000 characters of them, trimmed from the oldest end.
• Picking a thread in History now only opens it. It used to set the agent's Model to whatever that thread last ran on, changing the model you were about to run without telling you. The model stays the one you picked, and your next message runs it on the thread you opened.
• History rows no longer carry a tag naming the harness a thread ran on.
• The first message after a harness switch starts a new sandbox, so it takes longer than a follow-up. Messages after it are back to warm-sandbox speed.

A trigger wired into an agent used to hand it everything the service sent. A Slack message arrived as the whole delivery around it: verification tokens, authorization blocks, and a second copy of the text in Slack's own block format. The agent now reads a short account of what happened instead: the service and the kind of event, who it came from and where, the content itself, and the IDs it needs to reply in the right place.

• The triggers agents are usually wired to each write their own account of an event: Slack, Discord, Telegram, WhatsApp, inbound email, Gmail, Outlook, GitHub, GitLab, Linear, Jira, Notion, Trello, Asana, ClickUp, monday, Google Calendar, Forms, Sheets and Drive, Typeform, HubSpot, Shopify, Zendesk, Intercom, Stripe, Calendly, PagerDuty, Sentry, PostHog, Datadog, Honeycomb, Mailgun, Resend, Loops, schedules, and plain webhooks. Any other trigger gets the general shaping: the delivery wrapper opened, empty and internal fields dropped, and the rest held to a readable size.
• In the agent's chat, a turn that a trigger started is shown as the message that started it, laid out the way that app lays it out, under a caption naming the app, the node's label if you gave it one, and the trigger, for example Slack · Support channel · On app mention. Where a message view does not fit, such as a schedule tick, the account the agent read is shown with the event itself behind an Event payload toggle.
• The thread takes its name from the event, so the history list reads as the channel or the sender the message came from. Your standing instructions stay in the Message field, rather than appearing in the chat as a message you never sent.

Slack sends channel events only to apps that are members of the channel. A trigger on a channel NoClick was never invited to used to read Active and never fire. Registration now checks, and the trigger's Status field carries the verdict. When the app is not in the channel, the status turns amber and says so, with a Join #channel button next to it on a public channel. A private channel names the invite to run instead, since only a person can add an app to one.

• The node is marked incomplete on the canvas, counted by the incomplete pill at the bottom left, and named in the Setup tab. It never holds up a run you start yourself, because a manual run does not wait for Slack to deliver anything.
• Your Dashboard lists it under Needs you as not receiving events, with that same fix as the button on the row and Open trigger next to it. Open trigger now opens the workflow with the trigger selected and its settings showing.
• Ask the AI builder why a workflow is not running and it reads the same verdict, so it answers that the trigger is registered but not receiving events rather than calling it active.
• A check that cannot reach Slack changes nothing. A trigger is never called broken on a failed check.

Get Subreddit Posts reads Reddit directly now. Each post comes back with its full text and HTML, permalink, creation and edit times, score, total comment count, flair, media links, and the community's URL, description, and member count. It still needs no connected Reddit account, and it still runs on your credits.

• Set Fetch Comments to Yes for a sample of comments and their nested replies, sized by Comments per Post (1 to 100, 10 by default). num_comments stays the total Reddit reports for the post, and comment_count counts what actually came back.
• Time Period applies to top and controversial sorting on every read. Content Detail and Proxy Mode are gone, and steps you already saved keep running on the new reader.
• A read that fails or comes back incomplete fails the step, instead of reporting success with nothing in it.

The connect status for a workflow that hosts an MCP server now answers two questions separately instead of showing one spinner for both. A Link ready line names how many tools the link serves right now, with a few of them by name, read straight from your workflow with no client involved. Under it, the client line reads Waiting for your agent to connect until a client calls in, then Connected with the client's name, how many tools it was served, and which ones it can call.

• The waiting line names what makes that client connect. Most clients only write the config when you run their add command and open the connection when a session starts, so the usual fix is to start or restart the client rather than to wait.
• After 45 seconds the panel offers that client's own check command, such as claude mcp list or opencode mcp list. Those commands list the servers and connect at the same time, so running one turns the status green.
• The status also sits on the node. The Connect externally header of an MCP Server node reads Connected, with the client and the tool count, for anyone who opens the workflow later, so you do not have to run the Setup finale again to check.
• When NoClick cannot read handshakes for a moment, the panel says it cannot check rather than reporting that nothing has connected. The link keeps serving either way.
• The client setup guides are re-verified. Claude's custom connectors live under Customize → Connectors, Zed connects to a remote URL without a bridge, and VS Code adds the server with code --add-mcp.