github n8n-io/n8n n8n@2.5.0

pre-release14 hours ago

2.5.0 (2026-01-20)

Bug Fixes

  • ai-builder: Add guidance for referencing Webhook and AI Agent output (#24301) (206b3f3)
  • ai-builder: Correctly connect Mermaid sub-graphs (#24195) (ea6e742)
  • ai-builder: Disabling flaky E2E workflow builder test (#24341) (ca80b36)
  • ai-builder: Examples omitted in spec evals (#24171) (755a1f2)
  • ai-builder: Fix dirty state for nodes when importing workflows (#24165) (b7bebf9)
  • ai-builder: Long running prompts throw internal 401 error rather than expected complexity error (#24114) (71ca931)
  • Apply source control configuration changes to all multi main instances (#24038) (1d59287)
  • Bump dependencies (#24516) (502479c)
  • core: Add configurable HTTP timeout for AI nodes to prevent UND_ERR_HEADERS_TIMEOUT errors (#24292) (f6b0ff5)
  • core: Add LangSmith tracing metadata for V3 ToolsAgent (#24314) (a7e032c)
  • core: Allow resetting Error Workflow settings to default (#24299) (6252dd8)
  • core: Check user auth identities to know whether or not they can update their profile (#24524) (62afac3)
  • core: Fix command registration in modules (#24469) (b45d1f9)
  • core: Fix for execution history when flow includes wait node (#23146) (0e19f9c)
  • core: Fix workflow statistics constraint (#22620) (4aa9399)
  • core: Improve SSH host key verification for source control (#24408) (6a1bdd2)
  • core: Increase the model column length from 64 to 256 (#24332) (ac69fe5)
  • core: Insert a workflow index placeholder for workflows with no dependencies (#24190) (535caf4)
  • core: Preserve source overwrite in chained tool calls (#24226) (3e12788)
  • core: Put source control preferences behind auth (#24113) (8eb5e75)
  • core: Route both Python options to the native Python runner (#24511) (f72fdc0)
  • core: Support dots in usernames to connect to git repository in environments (#24481) (da868e3)
  • core: Use cache to enable MFA enforcement in multi main setup (#24261) (595ab25)
  • core: Use fsRealpath instead of resolve to get the real path (#24185) (fa6ad9c)
  • core: Worker status in multi-main only accessible to admin (#24548) (36cf19e)
  • editor: Adjust line height and vertical placement of icon in notification permission banner (#24549) (c7ce61b)
  • editor: Clear AI Builder chat when switching workflows (#24401) (0c721c3)
  • editor: Fix node issue not triggered on node disconnection (#24471) (57c2716)
  • editor: Fix parameter path in multipleValues=false fixedCollection (#24217) (053f2ca)
  • editor: Handle large values breaking data tables UI grid (#24470) (7b1e18a)
  • editor: Handle publishing conflict in webhook path at inaccessible workflow (#24327) (58a5e3e)
  • editor: Hold autosave when new template is imported until the user makes a change/run the workflow (#24334) (c337581)
  • editor: Hover main header tabs above evaluation page (#24265) (4bb3552)
  • editor: I18n support for Create a tag label (#20698) (08f0d94)
  • editor: Implement distance threshold for the Draggable component (#24490) (ca0ce66)
  • editor: Improve command bar loading state (#24441) (51140d0)
  • editor: Only open template setup modal on import (#24491) (0f21192)
  • editor: Only shift downstream nodes to the right of insertion point (#24052) (612947c)
  • editor: Prevent duplicate field_focus_placeholder_in_ndv telemetry event (#24383) (928a35f)
  • editor: Remove duplicate error toasts in ready-to-run workflow (#24335) (eb27530)
  • editor: Send builder telemetry event when session is loaded (#24297) (bfb11c0)
  • editor: Tweak publish tooltips for migrated workflows (no-changes) (#24296) (185b766)
  • editor: Update EULA acceptance modal and success messages for license activation (#23325) (ca0cddb)
  • Fix CLI import (#24215) (0c8b59a)
  • Fix issue preventing some community nodes from installing (#24448) (37b76f9)
  • Fix security issue in @rudderstack/rudder-sdk-node via major version upgrade from 2.1.4 to 3.0.0 (#24303) (bad85bb)
  • Form Node: Preserve CSS child combinator selectors in custom styling (#24106) (680f264)
  • Form trigger and Wait Form mode basic authentication fix for form POST request (#23795) (d2a0e86)
  • Harden Git node parameter handling (#24241) (503f299)
  • Implements SLSA L3 provenance and VEX attestation (#24411) (c589da7)
  • Kafka Trigger Node: Logs and better errors handling, rebalance timeout option (#24375) (3aa8535)
  • LmChatOpenAi Node: Fix tool calling with responses api against OpenAI-compatible backends (#24264) (5a4954f)
  • MCP Client Node: Make connection errors more clear (#24398) (4b0f472)
  • McpClientTool Node: Filter out tool arguments unless explicitly allowed (#24263) (389f9c4)
  • Microsoft Teams Node: Modify Microsoft Teams node default scopes (#23725) (aae8db4)
  • Oracle DB Node: Resolve SQL compatibility issues and correct outBind generation (#21489) (783d407)
  • Regenerate form webhook ids when pasting workflow data (#24192) (a19347a)
  • Set NODE_PATH for DHI images (#24525) (715a53c)
  • Skip SafeChain setup on Windows (#24576) (801a9ae)
  • Todoist Node: API migrating to v1 (#24240) (60f7697)
  • Unfork @n8n/vm2 (#24509) (71a6bd2)
  • Update base image node version (#24357) (81daa29)
  • Update urllib3 to 2.6.3 in task runner (#24325) (5567b91)
  • Use http proxy for source control git commands (#24104) (116b245)

Features

  • Add public API testing to e2e tests (#24194) (30decbb)
  • Add revokeUrl to status endpoint (#24253) (7d2e846)
  • ai-builder: Add auto-fix for missing AI node connections + prompt guide (#24519) (35f1999)
  • ai-builder: Add node recommendations for AI content generation (#24267) (59587ad)
  • ai-builder: Ask mode callout explaining its usage (#24406) (7bffd18)
  • ai-builder: Fitlter resource/operation in node-type definitions (#24078) (7c54ad3)
  • ai-builder: Fix AI builder prompt inconsistency about the Loop node (#24300) (f56bd1b)
  • ai-builder: Improve scheduling workflows generation (#24407) (9686582)
  • ai-builder: Make credentials validation minor instead of major (#24529) (70ef256)
  • ai-builder: Prefer ResourceLocator mode "list" for data storage nodes (#24181) (f140e66)
  • Allow editing / deleting project variables if you have the project scope (#24532) (d1647ca)
  • Allow skipping migration checks for imports (#24229) (934bd9a)
  • Binary data merge and expressions simplification (#23270) (0523cc3)
  • Chat Node: Regroup actions and add different response types (#23028) (0450be0)
  • ChromaDB Node: Add local chromadb support for complete local vector database (#19806) (5dd15e3)
  • core: Add Chat hub support for 'When Last Node Finishes' and 'Using Response Nodes' response modes (#24189) (10165c3)
  • core: Add user email parameter when activating EULA license (#23350) (3d71a6d)
  • core: Forbid updating archived workflows (#24528) (0bbb631)
  • core: Introduce long term workflow history trimming (#24382) (b7ea410)
  • core: Rename workflow history compaction env vars and short term compaction (#24384) (f85a238)
  • Data Table Node: Add Order By feature (#23677) (800075a)
  • editor: Add a clarifying callout to Merge Node schema view (#24435) (2ca8404)
  • editor: Add native browser notifications on AI workflow builder completion or waiting for input (#24224) (5ba0bf7)
  • editor: Add user_message_id to AI assistant feedback telemetry (#24279) (8e040f2)
  • editor: Autosave - Implement exponential backoff (#24464) (f0f4f37)
  • editor: Notify users on first prod error (#21764) (4e28e01)
  • editor: Show tooltips for canvas edge buttons (#24311) (bd111bf)
  • editor: Use number input component for time saved setting (#24331) (159c110)
  • GitHub Trigger Node: Add automatic webhook signature verification (#24203) (64c9148)
  • HTTP Request Node: Add option to disallow cross-origin credentials sharing (#24526) (5cd24a1)
  • LmChatOpenAi Node: Update default model to gpt-5-mini (#24342) (a44d4b1)
  • Update Posthog key name for execution logic v2 (#24288) (2e29bbe)

Performance Improvements

  • core: Disable data tables cleanup checks on workers (#24444) (3d90c72)
  • core: Disable unused modules on workers (#24449) (d7e5fea)
  • core: Reduce baseline memory usage by 33 MB for all instance types (#24107) (8ccad70)

Don't miss a new n8n release

NewReleases is sending notifications on new releases.