n8n release notes: editor and nodes https://docs.n8n.io/changelog/release-notes Feature-level updates from n8n releases: what shipped in the editor and the integration nodes. en Wed, 02 Sep 2026 13:01:00 GMT Add Contact Support link to Help menu for Cloud instances https://github.com/n8n-io/n8n/pull/37414 https://github.com/n8n-io/n8n/pull/37414 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 n8n Cloud instances now show a "Contact Support" item in the editor's Help menu, linking to support.n8n.io in a new tab. It appears between "Course" and "Report a bug" and is only visible on Cloud deployments, giving Cloud users quicker access to support resources. Self-hosted instances are unaffected. Align workflow and agent header menus https://github.com/n8n-io/n8n/pull/37315 https://github.com/n8n-io/n8n/pull/37315 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 The workflow and agent header menus have been aligned for consistency. The agent actions menu now sits next to the agent name, and version history has moved into that actions menu (disabled when no publish history exists). Both the workflow and agent menus now use the same dropdown component, with added icons and grouped items for a more consistent look. Existing import, export, favorite, and delete actions continue to work as before. Add floating Markdown editor toolbar https://github.com/n8n-io/n8n/pull/37293 https://github.com/n8n-io/n8n/pull/37293 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 The Markdown editor used for AI agent instructions now shows a floating formatting toolbar above selected text, instead of a toolbar that toggles between hidden and always visible. Selecting text reveals formatting options near the selection, and the toolbar disappears when the selection is cleared, giving a cleaner editing experience. Update the order of providers in the agent builder model picker https://github.com/n8n-io/n8n/pull/37143 https://github.com/n8n-io/n8n/pull/37143 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 The Agent Builder's model picker now orders providers to match the AI agent node's chat model picker: the currently selected provider appears first with a green 'Connected' checkmark, followed by n8n Connect-eligible providers grouped under an 'Included in n8n' header, then remaining providers, with aggregators like AWS Bedrock, OpenRouter, and Vercel AI Gateway listed last. A single divider separates the n8n Connect group from the rest, making it easier to find relevant providers. Confluence Node: Add Page Get Comments operation https://github.com/n8n-io/n8n/pull/37073 https://github.com/n8n-io/n8n/pull/37073 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 The Confluence node's Page resource now supports a Get Comments operation, listing footer comments on a page (one item per comment) via the Space + Page picker. It supports Return All/Limit with cursor pagination, Body Format (Storage, Atlas Doc Format, or Plain Text), and Sort By/Sort Direction options. Continue-on-fail is supported, and it returns no items for pages without comments. MongoDB Node: Batch update and find-and-update writes with bulkWrite https://github.com/n8n-io/n8n/pull/37035 https://github.com/n8n-io/n8n/pull/37035 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 The MongoDB node's Update and Find And Update operations now batch writes using bulkWrite instead of issuing one request per item, greatly speeding up large data ingests. This is available on new node version 1.5; existing workflows keep using their saved version, so behavior is unchanged unless you add a new MongoDB node. Output format and per-item results remain the same, but partial-failure scenarios (e.g. multi-collection or connection drop mid-write) may leave different write states than before. Databricks Node: Add user-delegated OAuth2 credential (authorization code + PKCE) https://github.com/n8n-io/n8n/pull/37007 https://github.com/n8n-io/n8n/pull/37007 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 The Databricks OAuth2 credential now supports user login via authorization code with PKCE, in addition to the existing service-principal client credentials flow, enabling per-user permissions and audit trails. Users can also customize OAuth scopes, with offline_access automatically preserved for token refresh on the user grant. Permission-denied errors now surface Databricks' own descriptive messages in executions, resource-locator dropdowns, and SQL failures, and trailing slashes in the configured host no longer break OAuth URLs. Existing service-principal credentials are unaffected. Add adaptive layout to ChatInput https://github.com/n8n-io/n8n/pull/36939 https://github.com/n8n-io/n8n/pull/36939 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 ChatInput now supports a new 'adaptive' layout that starts as a single compact row with an inline send/stop button and grows automatically as content is typed. It automatically falls back to the multiline layout if a custom button label or action slots are used, since adaptive only supports the default icon-only send/stop button. Gotify Node: Add Gotify extra notification settings https://github.com/n8n-io/n8n/pull/23915 https://github.com/n8n-io/n8n/pull/23915 Tue, 01 Sep 2026 00:00:00 GMT n8n 2.38.1 The Gotify node now supports additional message options based on Gotify's msgextras spec: Click URL (opens a link when the notification is clicked), Big Image URL (displays an image in the notification), and Intent URL (Android-only action after delivery). These are optional fields added to the message creation options and are only included in the request payload when set. Polish AI Assistant setup screen and connect modal https://github.com/n8n-io/n8n/pull/36795 https://github.com/n8n-io/n8n/pull/36795 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 Refined the self-hosted AI Assistant setup experience: the setup screen now has a clearer title and offers "Get started" or "Set up later in Settings" on first visit, switching to "Turn off for this instance" on return visits. The connect modal hides its step-indicator dots when only one setup step remains, and connection test failures now show an additional sanitized "Error details" line to aid troubleshooting. Add streaming to Message an Agent workflows https://github.com/n8n-io/n8n/pull/36734 https://github.com/n8n-io/n8n/pull/36734 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 Message an Agent workflows now support opt-in response streaming, with output streamed live through Chat and Webhook triggers. Active tool and skill calls made by the agent are shown as capability pills in the UI, giving users real-time visibility into what the agent is doing while it responds. Move AI Assistant connections from the sidebar to a "+" button in the text input https://github.com/n8n-io/n8n/pull/36585 https://github.com/n8n-io/n8n/pull/36585 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 AI Assistant connections (MCP servers, Computer Use, Browser Use) have moved from the sidebar Connections card into a new "+" button menu next to the chat input. The menu shows connected, connecting, and disconnected statuses, flags unexpected disconnects with a red indicator, and lets users reconnect Computer Use or Browser Use directly. Wording changed from "Tools" to "Connectors"/"Services". Show diff for closed review requests https://github.com/n8n-io/n8n/pull/36539 https://github.com/n8n-io/n8n/pull/36539 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 Closed workflow reviews now show a Changes tab when approved, comparing the version published before approval against the approved version, with each side labeled by its publish state. If a review closed without approval, the Changes tab now explains that there's nothing to compare instead of hiding the tab. Hide convert to sub-workflow when Execute Workflow nodes are excluded https://github.com/n8n-io/n8n/pull/36499 https://github.com/n8n-io/n8n/pull/36499 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 When an n8n instance excludes the Execute Workflow node type via NODES_EXCLUDE, the convert-to-sub-workflow action is now hidden from the canvas context menu, selection toolbar, group title bar, and the Alt+X shortcut, since conversion requires that node. Additionally, workflow settings' 'This workflow can be called by' option is locked to 'No other workflows' in this case. Default n8n behavior is unaffected since these nodes aren't excluded by default. Typeform Node: Allow custom OAuth2 scopes https://github.com/n8n-io/n8n/pull/36486 https://github.com/n8n-io/n8n/pull/36486 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 The Typeform OAuth2 credential now supports custom scopes. By default it still requests the same three scopes needed by the Typeform Trigger node, but users can enable a 'Custom Scopes' toggle to edit the requested scope string and access additional Typeform API permissions. Custom scopes are now preserved when reconnecting the account instead of resetting to defaults. Show review status as a badge on the workflows list https://github.com/n8n-io/n8n/pull/36476 https://github.com/n8n-io/n8n/pull/36476 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 Workflows with an open review now show a status badge on the Overview workflows list, next to the publish and project indicators. The badge links to the review inbox when you're allowed to open it, otherwise it displays as plain text; a tooltip explains the decision status. Closed reviews no longer show a badge. The canvas 'Approved' banner has been removed, since publishing is now handled via the regular Publish button. Requires workflow reviews (enterprise feature). Display success toast message on Browser Use connection https://github.com/n8n-io/n8n/pull/36474 https://github.com/n8n-io/n8n/pull/36474 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 When connecting Browser Use automatically (via credential setup or the sidebar connect button), n8n now shows a success toast and closes the connect modal once your browser is connected. The connection modal was also streamlined: the external-link icon on the "Open Browser Use Extension" button was removed since it now opens in a popup, and the waiting-for-connection indicator was removed. Microsoft SharePoint Node: Register version 2 and make it the default https://github.com/n8n-io/n8n/pull/36459 https://github.com/n8n-io/n8n/pull/36459 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 The Microsoft SharePoint node now defaults to version 2, which is built on the Microsoft Graph API and supports OAuth2 or Entra service principal authentication with site/list resource locators. New SharePoint nodes added to workflows will use version 2 by default. Existing workflows using version 1 continue to run unchanged, as version 1 and its credentials remain fully supported. Warn about self-escalation via Manage project roles scope https://github.com/n8n-io/n8n/pull/36458 https://github.com/n8n-io/n8n/pull/36458 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 The custom instance role editor now warns when a role has only the Manage project roles permission, noting this lets a holder edit the scopes of any project-type custom role, including one they themselves hold, potentially self-escalating permissions. This joins existing warnings for Manage all roles and Members: Manage. No authorization behavior changed; this is a UI clarity improvement for admins managing custom roles under the Custom Roles enterprise feature. Add stop execution button to the NDV https://github.com/n8n-io/n8n/pull/36444 https://github.com/n8n-io/n8n/pull/36444 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 You can now stop a node's test-step execution directly from the NDV header, without closing the panel. A stop button appears next to Execute step while a node runs; it cancels the running execution, or cancels the test webhook wait if the run is waiting for an incoming webhook call. This also fixes a race where a stop click issued right after starting could be dropped. Add Slack Agent view support https://github.com/n8n-io/n8n/pull/36352 https://github.com/n8n-io/n8n/pull/36352 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 n8n Agents connected to Slack now use Slack's newer Agent view for direct messages by default when configuring a new Slack app, including threaded replies. Existing Slack integrations continue using the legacy Assistant view to avoid disrupting current setups. This is a behind-the-scenes messaging experience change with no configuration required for new connections. Deprecate the "Any workflow" caller policy https://github.com/n8n-io/n8n/pull/36350 https://github.com/n8n-io/n8n/pull/36350 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 The workflow setting "This workflow can be called by → Any workflow" is now deprecated ahead of removal in n8n v3, since it lets any project on the instance call a sub-workflow, bypassing project permissions. The workflow settings dropdown now labels it deprecated and shows a warning when selected, prompting a switch to another policy. MCP tool and public API descriptions were also updated to discourage new use of this value. No behavior or enum values change yet. Add role-aware n8n credits top-up flow https://github.com/n8n-io/n8n/pull/35873 https://github.com/n8n-io/n8n/pull/35873 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 The n8n credits top-up flow is now role- and subscription-aware. Cloud instance owners on a paid plan are taken directly to the Cloud Admin Panel to top up. Members and admins instead see a modal explaining that only the instance owner can top up, with a list of connected services. Trial users see a modal noting a paid plan is required, and trial owners get an upgrade option. Anthropic Chat Model Node: Add prompt caching support https://github.com/n8n-io/n8n/pull/34482 https://github.com/n8n-io/n8n/pull/34482 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 The Anthropic Chat Model node (v1.6) now supports opt-in prompt caching. New Options fields let you enable caching and set a cache TTL of 5 minutes or 1 hour, allowing the system prompt, tool definitions, and conversation history to be reused across requests, reducing cost and latency for agents with large prompts or many tool calls. Caching is disabled by default, and existing workflows on earlier node versions are unaffected. Grist Node: Add upsert operation https://github.com/n8n-io/n8n/pull/18067 https://github.com/n8n-io/n8n/pull/18067 Tue, 25 Aug 2026 00:00:00 GMT n8n 2.37 The Grist node now supports a new Create or Update (upsert) operation, using Grist's built-in upsert API to create or update records in a single batched request. Users can define matching criteria fields, choose how to handle multiple matches (update first, none, or all), and populate fields via auto-mapping or manual definition, matching existing Create/Update UI patterns. Show scopes for managed OAuth credentials with N8N_MANAGED_OAUTH_SHOW_SCOPES https://github.com/n8n-io/n8n/pull/36397 https://github.com/n8n-io/n8n/pull/36397 Wed, 19 Aug 2026 00:00:00 GMT n8n 2.35.4 Added a new environment variable, N8N_MANAGED_OAUTH_SHOW_SCOPES, allowing admins to specify managed OAuth credential types (e.g. googleOAuth2Api) that should still expose the scope field for customization, even when credentials are overwritten via environment variables. Previously, scope fields were always hidden for managed credentials with overwrites, breaking workflows that relied on customizing scopes. Improve agent session tool call feedback https://github.com/n8n-io/n8n/pull/36346 https://github.com/n8n-io/n8n/pull/36346 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Agent session timelines now separate tool calls, approval requests, and approval responses into distinct events for clearer tracking. Preview chat shows expandable resolved tool call details and distinguishes declined approvals from failed calls. Dismissed MCP connection warnings now stay hidden for the rest of the current preview session, reducing repetitive alerts. Simplify user flow for Browser Use connection https://github.com/n8n-io/n8n/pull/36316 https://github.com/n8n-io/n8n/pull/36316 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Connecting your browser to n8n's Browser Use AI Assistant is now simpler: clicking connect automatically opens the extension's connect view in a popup, which closes itself once connected. If the popup is closed or times out after 15 seconds, a Try again button lets you refocus it. The manual Open Browser Use extension button now also opens a popup instead of a new tab. Adapt the OAuth consent screen for first-party clients https://github.com/n8n-io/n8n/pull/36261 https://github.com/n8n-io/n8n/pull/36261 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 The OAuth consent screen now recognizes first-party clients, such as Form Trigger nodes using n8n login authentication. Instead of the generic third-party (MCP) layout, users see the form's icon, a tailored heading and description, and no trust checkbox or redirect-URI warning, letting them grant access immediately. Third-party MCP consent flows are unchanged. Improve Browser Use connect conditions https://github.com/n8n-io/n8n/pull/36215 https://github.com/n8n-io/n8n/pull/36215 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Browser Use setup in the AI Assistant now better handles unsupported environments. On mobile and tablet devices, where browser extensions aren't possible, Browser Use and automatic credential setup are hidden entirely. On non-mobile devices with non-Chromium browsers, the option still appears but the connect modal now clearly explains that only Chrome or Chromium-based browsers, like Edge or Brave, are supported. Telemetry now also records browser support status to help track success rates. Add preview chat to SessionTimeline view https://github.com/n8n-io/n8n/pull/36128 https://github.com/n8n-io/n8n/pull/36128 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Agent builder session views now include a preview chat you can dock beside the session timeline or open full screen, so you can inspect a trace and continue the conversation without switching pages. You can switch sessions and control the preview directly from the dock, open a preview in a new tab, start a new session from the agent list, and your preferred preview layout is remembered across the agent builder and session views. Add snapshot option to every browser use interaction tool https://github.com/n8n-io/n8n/pull/36111 https://github.com/n8n-io/n8n/pull/36111 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Browser Use interaction tools (click, type, select, drag, hover, press, scroll, upload, dialog) now accept an optional snapshot parameter to return an accessibility snapshot directly in the response, either interactive or non-interactive. This removes the need for a separate snapshot call after each action, roughly halving the number of tool calls needed for browser automation. Let users update review descriptions https://github.com/n8n-io/n8n/pull/36069 https://github.com/n8n-io/n8n/pull/36069 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 When updating a workflow review with a newly pinned version, users can now edit the review description as part of the update flow. Review descriptions are preserved through the update, and are only visible to users with publish permission. The update dialog is more resilient to stale review state and description prefill failures. MCP Server Trigger Node: Add server instructions option https://github.com/n8n-io/n8n/pull/35945 https://github.com/n8n-io/n8n/pull/35945 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 The MCP Server Trigger node now has an optional Instructions field. Text entered here is sent to connecting MCP clients as part of the initialize response, letting you provide cross-tool guidance such as tool-choice rules or multi-step workflows that don't belong in any single tool's description. Clients that support this MCP feature (e.g. Claude Code) surface it in their system prompt automatically; others ignore it. The field defaults to empty, so existing workflows are unaffected. Update NumberInput styling https://github.com/n8n-io/n8n/pull/35937 https://github.com/n8n-io/n8n/pull/35937 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 The internal NumberInput design system component now visually matches the standard text input, with consistent sizing, borders, and focus/hover states, plus icon-based increment/decrement buttons with accessible labels and focus-to-select-all behavior. Currently this component is only used on the Agent Builder Settings page, affecting fields like Concurrency, Max iterations, Max uses, and Max parallel sub-agents. MiniMax Node: Add M3 and H3 model support https://github.com/n8n-io/n8n/pull/35935 https://github.com/n8n-io/n8n/pull/35935 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 The MiniMax node now supports the MiniMax-M3 language model, set as the new default, alongside the existing model lineup. It also adds MiniMax-H3 support for text-to-video and image-to-video generation via MiniMax's V2 video API, with dedicated H3 parameters like duration, resolution, and aspect ratio, while previous models and workflows using the V1 API continue to work unchanged. Display authors on the review details view https://github.com/n8n-io/n8n/pull/35897 https://github.com/n8n-io/n8n/pull/35897 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Workflow review requests now show the workflow's authors alongside the requester and reviewers. The review details view lists a 'Requested by' entry, other authors, and reviewers, while the inbox sidebar cards display an author avatar stack instead of full avatars. Layout issues causing inbox cards to grow beyond their intended width were also fixed. This is part of the enterprise workflow reviews module. Link review success toasts https://github.com/n8n-io/n8n/pull/35888 https://github.com/n8n-io/n8n/pull/35888 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Success toasts shown after submitting a workflow for review, or after submitting an updated version to an existing review, now include an 'Open review' link. Clicking it navigates directly to the corresponding review, making it faster to check on submitted or updated review requests. Add docked agent preview chat https://github.com/n8n-io/n8n/pull/35773 https://github.com/n8n-io/n8n/pull/35773 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 The agent builder now docks a preview chat panel alongside the editor instead of switching to a separate full-page view, letting you test agent changes without losing your place. Session traces still open full-page with improved timeline navigation, and a new three-column layout supports Instance AI agent workflows. Sending sessions to Instance AI and cost data in session traces were also improved. Schedule Trigger Node: Add a per-node missed execution grace period https://github.com/n8n-io/n8n/pull/35771 https://github.com/n8n-io/n8n/pull/35771 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Schedule Trigger nodes now have their own 'Missed Execution Grace Period (Seconds)' setting, letting you configure how late a run can start before it's considered missed on a per-node basis, instead of relying only on the instance-wide N8N_SCHEDULER_MISFIRE_GRACE setting. Leaving it at 0 keeps the instance default. Invalid values fall back to the instance setting with a logged warning. Available on Schedule Trigger nodes at version 1.4 and above. Confluence Node: Add Confluence Cloud OAuth2 credential https://github.com/n8n-io/n8n/pull/35732 https://github.com/n8n-io/n8n/pull/35732 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 Added a new Confluence Cloud OAuth2 credential type, laying groundwork for an upcoming native Confluence node. It shares a new Atlassian OAuth2 base credential with Jira, supporting a Site URL field and customizable scopes. Jira's OAuth2 credential was refactored onto this shared base, fixing site matching issues with scheme-less or trailing-slash domains; existing Jira credentials continue to work unchanged. Schedule Trigger Node: Add "If Execution Is Missed" option https://github.com/n8n-io/n8n/pull/35626 https://github.com/n8n-io/n8n/pull/35626 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 The Schedule Trigger node now has an "If Execution Is Missed" option (node version 1.4, in the Settings tab) letting you choose whether an occurrence that comes due while the instance is down is skipped or run once instance comes back up. Existing workflows default to skipping missed runs unless coalesce is explicitly chosen. Also fixed a bug where increasing the grace period could incorrectly revive already-expired occurrences. Discord Node: Add member moderation actions https://github.com/n8n-io/n8n/pull/33486 https://github.com/n8n-io/n8n/pull/33486 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 The Discord node's Member resource now supports four moderation actions: Ban, Unban, Kick, and Timeout. Ban includes an optional message-cleanup window, and Timeout offers duration presets up to Discord's 28-day maximum or removal. All actions support a preset or custom reason, recorded in the server's audit log, enabling workflows to moderate flagged accounts directly, e.g. trigger → AI scorer → ban/kick/timeout. lmChatOpenRouter Node: Add provider routing options https://github.com/n8n-io/n8n/pull/24822 https://github.com/n8n-io/n8n/pull/24822 Tue, 18 Aug 2026 00:00:00 GMT n8n 2.36 The OpenRouter Chat Model node now supports provider routing options, letting you control which sub-providers handle requests. New settings under Options > Provider Routing include order, allowFallbacks, requireParameters, dataCollection, zdr, only, ignore, and sort, giving finer control over provider selection, fallback behavior, data retention requirements, and cost/performance sorting. GitHub Node: Add timeout option to Dispatch and Wait for Completion https://github.com/n8n-io/n8n/pull/35837 https://github.com/n8n-io/n8n/pull/35837 Tue, 11 Aug 2026 00:00:00 GMT n8n 2.35 The GitHub node's Dispatch and Wait for Completion operation now supports a Limit Wait Time option, letting you cap how long the execution waits for the dispatched GitHub Actions workflow to call back, either after a time interval or until a specified date and time. Previously the execution waited indefinitely if the callback never arrived. If the limit is not set, existing behavior is unchanged; on timeout the node resumes silently with its input data instead of the callback payload. Kafka Node: Add compression options to version 2 https://github.com/n8n-io/n8n/pull/35727 https://github.com/n8n-io/n8n/pull/35727 Tue, 11 Aug 2026 00:00:00 GMT n8n 2.35 The Kafka node (version 2) now includes a Compression option supporting None, GZIP, Snappy, LZ4, and Zstd, letting users choose how messages are compressed before sending. Previously, only GZIP was available via a toggle in version 1, since the older library lacked native support for other codecs. Version 1 behavior is unchanged. Note that version 1 of the Kafka Trigger cannot decode Snappy, LZ4, or Zstd, so use GZIP or None if version 1 triggers consume the topic. Merge Node: Add "Append" and "Combine" search aliases https://github.com/n8n-io/n8n/pull/35709 https://github.com/n8n-io/n8n/pull/35709 Tue, 11 Aug 2026 00:00:00 GMT n8n 2.35 You can now find the Merge node in the nodes panel by searching for "Append" or "Combine", in addition to its existing search terms. This makes the node easier to discover when you're looking for ways to combine or append data streams. Allow adding a workflow version description when submitting for review https://github.com/n8n-io/n8n/pull/35640 https://github.com/n8n-io/n8n/pull/35640 Tue, 11 Aug 2026 00:00:00 GMT n8n 2.35 You can now add a version description when submitting a workflow for review, matching the existing publish flow. Submit for review is now a two-step dialog for naming and describing the version before adding review details, and updating an existing review request now includes a description field. Empty values clear the description; omitting it leaves it unchanged. Add review detail metadata https://github.com/n8n-io/n8n/pull/35638 https://github.com/n8n-io/n8n/pull/35638 Tue, 11 Aug 2026 00:00:00 GMT n8n 2.35 Workflow Reviews inbox now shows a metadata sidebar on the review detail view, with cards for status, reviewers (avatar and name), and workflow changes. A status dot appears next to the review title, and clicking a workflow in the Changes card jumps to its diff section. Long names truncate gracefully, and empty reviewer or workflow lists show fallback messaging. Name the pinned workflow version when submitting for review https://github.com/n8n-io/n8n/pull/35566 https://github.com/n8n-io/n8n/pull/35566 Tue, 11 Aug 2026 00:00:00 GMT n8n 2.35 When submitting a workflow for review (or re-pinning an updated version), you now provide a required version name, prefilled from the version's existing name or a generated default. This prevents the pinned review version from being silently pruned from workflow history, which previously could null out the review's pinned version and break auto-publish on approval.