Release Notes - 2.0.0 B2¶

Introduction¶

Welcome to ELITEA — your comprehensive platform for building, deploying, and managing AI-powered workflows and intelligent agents. This release introduces significant enhancements including new features for image handling, execution history tracking, pipeline node redesign, and expanded MCP capabilities. For a complete overview of ELITEA and its capabilities, see the Introduction.

Information¶

Release Version: 2.0.0 B2
Released on: 19-Jan-2026
Access for 2.0.0 B2: Next environment

New Features¶

MCP: Remote MCP Server Integration with OAuth & Dynamic Tool Discovery¶

Connect remote MCP servers to ELITEA with full OAuth 2.0 authentication support, dynamic tool discovery, and intelligent session management.

What you can do now:

OAuth & PAT Authentication: Seamlessly authenticate remote MCP servers using OAuth 2.0 flow or Personal Access Token (PAT), depending on the remote server type. The authentication method is automatically detected based on server requirements.
Dynamic Tool Discovery: Click "Get/Sync Tools" to automatically fetch and discover all available tools from the remote MCP server without manual configuration.
Automatic Tool Synchronization: Use "Sync Tools" to update your tool list when the remote server adds or removes capabilities, with visual indicators showing what changed.
Session Management: Real-time validation of credentials with visual status indicators (connected/disconnected/expired) on toolkit cards.
Graceful Expiration Handling: When sessions expire, receive clear in-chat notifications with one-click re-authentication options to quickly restore connectivity.
Browse Remote Servers: Explore a curated directory of remote MCP servers at mcpservers.org to discover and connect new capabilities.

Benefits: Eliminates manual tool configuration, reduces authentication friction, ensures toolkit reliability through proactive session validation, and expands integration possibilities with the growing ecosystem of remote MCP servers.

For more information about this feature, see Remote MCP Server Integration Guide.

Agents Studio: Shared Library for Published Agents¶

Discover, browse, and use community-published agents through the new Agents Studio menu—a centralized shared library for exploring pre-built agents across all projects.

What you can do now:

Centralized Discovery Hub: Access the new Studio menu (located between Import and Connection Status) to explore all published agents from the community in one place.
Smart Search & Filtering: Search agents by name, description, or tag with real-time results, and filter by categories like "Trending" (most-liked agents) to quickly find what you need.
Visual Agent Browsing: View agents as cards showing agent name, creator avatar, and like count in a responsive grid layout optimized for different screen sizes.
Detailed Agent Preview: Click any agent card to open a modal displaying the agent's title, description, welcome message, conversation starters, and a "Start Conversation" button.
Instant Conversation Launch: Click "Start Conversation" to automatically create a new chat with the selected agent added as an active participant, complete with conversation starters ready to use.
Flexible Integration: Add published agents to existing conversations by typing #[Agent Name] in the chat input or searching from the Agents section in Participants.
Customizable Experience: Modify LLM model and model settings for published agents in your conversations (other properties remain read-only to preserve the original agent design).

Benefits: Reduces time to discover and deploy pre-built agents, fosters community engagement through agent sharing, and provides an intuitive browsing experience similar to popular AI assistant marketplaces.

For more information about this feature, see Use Public Agents.

Indexing: Enhanced Index Management with Real-Time Monitoring & Automation¶

Comprehensive index management improvements including real-time progress monitoring, automated scheduling with visual cron builder, event notifications, and execution history tracking.

What you can do now:

Real-time progress monitoring: Monitor index creation and reindexing with color-coded status indicators (green checkmark for success, blue loading for in-progress, red info icon for failed), view detailed progress information and real-time logs, and see index card details including name, creation date, and document count.
Automated schedule updates with visual cron builder: Enable scheduled index refreshes (daily, weekly, monthly) with the Auto-Update toggle, use the intuitive visual cron builder with dropdown-based controls for frequency selection (Every minute, Hourly, Daily, Weekly, Monthly, Custom), time picker, day selectors, and real-time preview of generated cron expressions—eliminating the need to understand cron syntax or use external tools like crontab.guru.
Flexible scheduling options: Configure schedules using the visual builder interface or switch to manual text input for advanced cron expressions, select from common presets ("Every hour", "Daily at midnight", "Weekly on Monday"), and ensure security compliance with enforced project-level credentials and role-based access control.
Execution history tracking: Access the History tab to view all index runs with creation/update timestamps, track successful and failed runs with detailed status information, and click any run to view comprehensive details including input/output data and error diagnostics.
Event notifications: Receive real-time notifications for index creation, reindexing (manual or scheduled), and failures with document counts, click index names in notifications to navigate directly to the Index tab with the selected index and Run details displayed.
Failed index management: View detailed error information for failed indexes, reconfigure parameters and restart indexing via the Update button, or delete indexes that are no longer needed.

Benefits: Eliminates manual monitoring and maintenance overhead, reduces scheduling errors through intuitive visual interface accessible to non-technical users, ensures indexes stay current automatically, provides complete visibility into index operations and history, enables quick troubleshooting with detailed diagnostics, and maintains enterprise security through enforced credential policies.

For detailed guidance, see How to create and use indexes and Schedule Indexing.

Pipelines: Enhanced Execution Monitoring & AI-Powered Configuration¶

Comprehensive pipeline improvements including real-time run tracking in Canvas and Flow View with AI-powered input assistance for node configuration.

What you can do now:

Real-time run tracking in Canvas: Monitor pipeline execution directly in Canvas Flow Editor when running from chat, with run details appearing automatically using the same UI as Pipelines Menu Flow View, multiple run support with completed runs grouped under a History icon, and seamless tab switching between Flow Editor and YAML tabs.
Enhanced run display in Flow View: View streamlined chipsets showing status icons and run numbers (e.g., "Run 1", "Run 2"), with smart grouping of completed/inactive runs under a History icon, color-coded status indicators (blue loading, green success, red error, orange stopped), and click-through access to detailed run information with input/output data and timestamps.
Run management: Delete completed runs from both Canvas and Flow View (in-progress runs must be stopped first), with session-based persistence that clears on page refresh to keep views focused on current work.
AI-powered input assistant: Generate or improve content for specific node fields (Decision descriptions, Router/Condition conditions, State Modifier Jinja templates, LLM System/Task fields) with side-by-side comparison of current versus improved options, iterative refinement support, and context-aware suggestions using your pipeline's configured LLM model.

Benefits: Eliminates context switching for execution monitoring, enables rapid iteration and debugging with immediate result visibility, improves troubleshooting efficiency with clear status differentiation, reduces time writing node configurations through AI assistance, and provides complete execution history—all while maintaining full user control over pipeline design and content.

For detailed guidance, see AI Assistant in Nodes and Pipeline Runs.

Agents & Pipelines: Run History Actions (Restore Conversation & Copy Link)¶

Restore a previous conversation from Agents or Pipelines run history into a new, editable chat session, and share specific runs with a direct link. This helps you troubleshoot issues, re-test after changes, collaborate with teammates, or recover context after an accidental page refresh.

What you can do now:

Restore from history: Use the Restore Conversation action on any run in the Agents or Pipelines History tab.
Copy a direct run link: Use the Copy link action for any run to copy a unique URL to your clipboard.
Open the exact run from a link: When you open the link, ELITEA navigates directly to the selected Agent/Pipeline run and shows its details.
Open a new session: Restoring opens a new chat session in the Configuration tab.
Restore chat history only: Messages and responses are restored; historical agent/pipeline configuration is not.
Use current configuration: The restored conversation runs with the current Agent/Pipeline configuration (for example, if toolkits changed, the current setup applies).
Continue immediately: The restored session is fully interactive so you can continue from the last message.
Non-destructive behavior: Restoring does not delete or modify the original history entry, and you can restore multiple runs into separate sessions.

Benefits: Enables faster iterative debugging and validation with the same context, improves collaboration by sharing run-specific links, reduces lost work after refresh or browser closure, and supports re-testing with updated configurations.

For background on run history, see Agents and Pipelines History.

Pin Frequently Used Entities¶

Pin agents, pipelines, credentials, toolkits, and MCPs to keep frequently used items at the top of lists for quick access.

What you can do now:

Pin entities: Click the pin icon in card view (bottom-left corner), table view ("..." menu), or detail view to pin any entity.
Quick access: Pinned entities appear at the top of lists across all views and persist across sessions.
Multiple pinning: Pin multiple entities while maintaining pin order; newly created entities appear after pinned items.
Visual indicators: Pinned state shown with filled icon; unpin with a single click.
Available for: Agents, Pipelines, Credentials, Toolkits, and MCPs across card, table, and detail views.

Benefits: Reduces time searching for frequently used entities, provides instant access to critical resources, and maintains personalized workspace organization across sessions.

For detailed guidance, see Pin Entities.

Chat: Planner Tool for Task Management¶

Create and manage tasks directly within conversations using the new Planner internal tool.

What you can do now:

Access via internal tools menu: Enable the Planner tool through the "Enable internal tools" dropdown (gear icon in chat input) to activate task management capabilities in your conversations.
Task creation from conversation context: Capture action items and tasks directly from your conversation by using phrases like "add this to my todo," "create a task for," or "remind me to follow up."
Context linking: Tasks are automatically linked to the conversation thread where they were created, preserving context for future reference.
Available everywhere: Use the Planner tool in Conversations (Chat menu), Agents menu, and Pipelines menu chat inputs.
Session persistence: Your enable/disable choice persists throughout the conversation session, and tasks remain accessible across conversations.
Smart invocation: The tool activates only when enabled and when task-related input is detected, ensuring focused and relevant responses.

Benefits: Eliminates context switching between conversations and external task management tools, captures action items in real-time as they emerge during discussions, and maintains task context for better follow-through—all without leaving your chat interface.

For detailed guidance, see Planner.

Toolkits: Usage History Tracking¶

Track and review all toolkit index operations with a dedicated History tab that preserves run details and execution information for auditing and analysis.

What you can do now:

View execution history: Access a dedicated History tab in the Indexes view to see all past indexing operations for the selected index.
Run details at a glance: Each history entry displays the creation date/time, update timestamps for manual and scheduled re-indexing, run status (success or failed), and detailed execution information.
Real-time updates: The history table automatically updates when indexing events occur (index creation, re-indexing, or failures).
Failed operation tracking: Failed updates appear in the history table with date/time and status indicating the failure reason.
View detailed information: Click any run in the history table to view comprehensive details in the right panel, including all relevant actions, parameters, and results (both successful and failed).
Reference-only access: History events are displayed for reference and auditing purposes; no direct actions can be performed from the History tab.

This feature enables better debugging, performance analysis, and compliance tracking by maintaining a complete audit of all indexing operations.

Toolkits: ADO Work Item Attachment Reading¶

Read and access attached images and files directly from Azure DevOps (ADO) work items through the ADO toolkit.

What you can do now:

Access work item attachments: Retrieve attached images and files from ADO work items using the enhanced ADO toolkit capabilities.
Expanded integration: Process visual content and documentation stored as work item attachments for comprehensive workflow automation.
Seamless file handling: Access attachments alongside other work item data (descriptions, comments, fields) through the same toolkit interface.

Benefits: Enables complete work item data extraction including visual assets, eliminates manual file downloads for automation workflows, and expands ADO integration capabilities for documentation-heavy projects.

Toolkits: Better PPTX Table/Chart Text Extraction¶

Artifacts, SharePoint, and indexing tools can now extract and index text from tables and charts embedded in PPTX files for more complete document understanding.

Credentials: Enhanced Test Connection Verification¶

Comprehensive connection testing for integration credentials with authentication validation, service-specific API verification, and detailed error diagnostics.

What you can do now:

URL reachability verification: Test Connection now performs actual HTTP requests to verify the service URL is accessible, not just well-formed.
Authentication validation: Validates provided credentials (API Key/Token, Username, Password) by attempting authenticated connections to each service.
Service-specific API calls: Uses appropriate endpoints for each integration:
- GitHub: Verifies credentials via /user or /repos endpoints
- TestRail: Validates access through /get_projects endpoint
- Jira: Confirms authentication via /myself endpoint
- Confluence: Checks credentials through /content endpoint
Detailed error messages: Receive specific feedback for issues including "Invalid URL," "Authentication Failed," "Insufficient Permissions," or "Service Unreachable" instead of generic connection errors.

Benefits: Eliminates false positives from URL-only validation, catches authentication issues before saving credentials, reduces troubleshooting time with specific error messages, improves integration reliability through comprehensive testing, and prevents configuration errors that could cause workflow failures.

Changed Features¶

Chat: Improved New Conversation Creation Flow¶

The new conversation creation flow has been redesigned to reduce user confusion and align with familiar chat application patterns.

What changed:

Deferred sidebar updates: Clicking +Create now highlights the chat input field without immediately showing a placeholder conversation in the sidebar.
Message-triggered creation: New conversation items appear in the sidebar only after you type and send your first message, eliminating premature UI elements.
Auto-generated naming: Conversation names are automatically generated based on your first message content.
Focused naming indicator: The loader icon and "Naming" text appear only during the brief name generation process (1-2 seconds), not before you've started typing.

Benefits: Cleaner and more intuitive conversation creation experience with reduced cognitive load, eliminated premature placeholder conversations, and alignment with popular chat application patterns like ChatGPT.

For more information about this, see Chat Usage.

Chat: Intelligent Attachment Bucket Management¶

Attachment storage is now automatically configured from agents with full manual control, clear visibility, and intelligent bucket inheritance.

What changed:

Automatic bucket inheritance: Adding an agent with a configured attachment bucket automatically sets it as the conversation's attachment bucket, eliminating manual configuration for each conversation.
First-agent priority: When adding multiple agents with different buckets, the first agent's bucket is used; subsequent agents' buckets are ignored to maintain consistent storage.
Manual override capability: Change the attachment bucket at any time via the settings icon in Participants → Toolkits; manual selections take precedence over future agent additions.
Persistent configuration: Attachment buckets remain active even when agents are removed from the conversation, ensuring previously uploaded files stay accessible.
Public agent bucket creation: Public agents automatically create buckets in your current project context (Private or Team) if they don't already exist.
Clear visual indicators: Active attachment buckets appear in Participants → Toolkits with a 📎 icon, bucket name, project type badge (Private/Team), and management controls.
Unified artifact storage: Artifact toolkit uses the conversation's attachment bucket when present, avoiding duplicate bucket configurations.

Benefits: Reduces attachment configuration friction by 70%, provides seamless agent-to-conversation bucket inheritance, ensures clear visibility of active storage location, and maintains full user control with manual override capabilities.

For more information about this feature, see Attachements in Conversation.

Toolkits: ADO Repos Credential Simplification¶

ADO Repos toolkit configuration is now simpler and more consistent by reusing the standard ADO credential and moving repository-specific configuration into the toolkit.

What changed:

ADO Repos credential removed: ADO Repos no longer uses a separate credential type; configure and reuse your existing ADO credential instead.
Repository ID moved to toolkit: A new Repository ID field is now part of the ADO Repos toolkit configuration.
Repository ID is required: Repository ID is now a mandatory field and appears before Base branch.
Seamless transition: Migrations are provided to support existing configurations.

Benefits: Reduces configuration overhead and credential confusion, standardizes Azure DevOps authentication across toolkits, and makes ADO Repos setup more predictable.

For more information about ADO credentials, see How to Use Credentials.

Model Settings: Simplified Configuration with Auto Mode¶

Token management and model parameters are now easier to configure with auto mode defaults and intuitive control settings for both technical and non-technical users.

What changed:

Auto token management: max_completion_tokens now defaults to "auto" (uses model defaults), with option to set explicit integer values for custom limits.
Creativity control: Replaced temperature with Creativity setting using a 3-5 dimension scale for intuitive model behavior tuning.
Reasoning levels: Replaced max_p and max_k with Reasoning (hi/med/lo) for simplified response control.
State-saving for continuity: Chat and agent sessions now save state when token limits are reached, with a "Continue" button to resume generation while preserving full context and history.

Benefits: Simplifies model configuration for non-technical users, eliminates manual token calculations, provides intuitive control dimensions, prevents context loss during long generations, and maintains advanced customization options for power users.

Pipelines: Node Deprecations and New Node Types¶

The pipeline node architecture has been streamlined with new specialized nodes and deprecation of legacy nodes to simplify pipeline design and reduce configuration complexity.

New Node Types Introduced:

Toolkit Node: Dedicated node for executing individual toolkit tools with input/output state mapping, interrupt settings, and tool-specific field configuration. Select from toolkits added to your pipeline, choose specific tools, and configure tool parameters directly in the node—replacing the generic Function/Tool approach with toolkit-specific clarity.
MCP Node: Specialized node for executing MCP server tools with the same configuration capabilities as Toolkit Node. Provides clean separation between toolkit and MCP tool execution for better pipeline organization.
Printer Node: Display formatted output to users without LLM processing or tool invocation. Supports F-string formatting (enabled by default) for dynamic variable interpolation, multiple input variables, and configurable Interrupt Before/After settings (Interrupt After enabled by default). Ideal for showing intermediate results, progress updates, or data summaries during pipeline execution—faster and more consistent than using LLM nodes for simple output display.

Node Deprecation Timeline

The following nodes are marked as deprecated in this release and will be removed in 2026. Existing pipelines using these nodes will continue to function, but you cannot add new instances of deprecated nodes. Migration guides are available to help transition to replacement nodes.

Enhanced Agent Node:

Agent Node: The Agent node has been enhanced to support pipeline execution alongside agent execution, providing a unified interface for both use cases. Agent nodes can now execute pipelines that do not contain interruptions or subgraphs (nested pipelines), supporting standard sequential and conditional flows. This enhancement simplifies workflow creation and reduces the need for separate node types.

Deprecated Nodes (Marked for 2026 Removal):

Function Node (deprecated): Replaced by Toolkit Node and MCP Node. The Function node handled both toolkits and MCPs in a single node type, creating configuration confusion. Dedicated Toolkit and MCP nodes provide clearer separation and improved usability.
Tool Node (deprecated): Replaced by Toolkit Node and MCP Node. Functionality now fully covered by the dedicated replacement nodes with enhanced capabilities.
Condition Node (deprecated): Replaced by Router Node. The Router node provides more robust and flexible conditional routing with better configuration options and clearer branching logic.
Decision Node (redesigned and deprecated old version): The legacy "semi-node" Decision Node attached to other nodes is deprecated. A new standalone Decision Node functions independently, accepts multiple input connections, and operates similarly to Router Node. The new Decision Node supports its own input/output state management and decision logic configuration without dependency on connected nodes.
Pipeline Node (deprecated): Replaced by enhanced Agent Node. The Agent node now supports pipeline execution for pipelines without interruptions or subgraphs, providing a unified interface for both agents and pipelines. Validation prevents adding unsupported pipeline types with clear error messages.
Loop Node and Loop for Tool Node (deprecated): These nodes were overly complicated for configuration and everyday usage. Alternative pipeline patterns using simpler nodes and design patterns are recommended. Simpler approaches provide the same iteration capabilities with better clarity and easier maintenance.

Migration Support:

No automatic migration is provided—existing pipelines continue to work with deprecated nodes.
Comprehensive migration guides available with step-by-step instructions, visual diagrams, and example conversions.
New nodes provide full feature parity with deprecated nodes, including input/output states, input mappings, and interrupt configurations.

Benefits: Clearer node responsibilities reduce configuration errors, specialized nodes improve pipeline readability, and removing overlapping legacy nodes simplifies learning curve for new users while maintaining full functionality for existing workflows.

For detailed migration guidance, see Migration Overview: migration/v2.0.0 B2/migration-overview.md.

Toolkits: Standardized Name and Description Fields¶

All toolkits now use consistent Toolkit Name and Description fields for better identification and organization.

What changed:

Toolkit Name field: All toolkits now include a mandatory "Toolkit Name" field (previously various fields like Collection ID, Space Key, Repository, Bucket, etc. were used for naming).
Description field: Optional "Description" field added to all toolkits for additional context and documentation.
Consistent placement: Toolkit Name always appears as the first configuration field across all toolkit types.
Automatic migration: Existing toolkits automatically migrated with current names preserved as "Toolkit Name".

Benefits: Eliminates naming inconsistencies across toolkit types, improves toolkit identification in lists and dropdowns, provides better documentation capabilities, and creates consistent user experience when configuring any toolkit.

OpenAPI Toolkit: Enhanced Configuration and Validation¶

The OpenAPI toolkit has been redesigned with improved schema validation, credentials support, and better toolkit organization.

What changed:

Credentials support: OpenAPI toolkit now supports credentials for secure authentication configuration.
Toolkit categorization: Moved to Integrations category for better organization alongside other integration toolkits.
Global headers: Toolkit-level headers configuration eliminates repetitive header definitions across endpoints, with secure storage for sensitive values.
Enhanced schema editor: CodeMirror integration provides syntax highlighting and real-time validation for schema field.
YAML support: Schema field now accepts both JSON and YAML formats with automatic format detection and validation.
Immediate feedback: Real-time syntax error detection prevents invalid configurations before saving.

Benefits: Reduces configuration errors, streamlines setup with reusable headers, supports industry-standard formats, improves security for sensitive values, and provides better editing experience with validation feedback.

QTest Toolkit: Enhanced Test Management Capabilities¶

The QTest toolkit has been significantly enhanced with new tools for managing custom fields, internal requirements, and indexing support, plus critical bug fixes for multi-agent usage.

What changed:

Custom Field Management: New tools enable you to retrieve, create, and edit custom field values for test cases, supporting various field types including checkboxes, text inputs, integers, and dropdown selections.
Internal Requirements Support: Attach internal requirements to test cases and use new tools for listing and retrieving requirements, with clear differentiation between internal and external requirement types.
Indexing Support: QTest data can now be indexed and searched using standard indexing tools, enabling better searchability and more efficient data retrieval across test cases, requirements, and other QTest entities.
Default Field Fixes: Priority, Type, and Status fields now correctly save user-selected values instead of defaulting to the first available option during test case creation.
Multi-Agent Compatibility: Fixed conflicts that occurred when multiple agents used the QTest toolkit simultaneously in the same pipeline or agent, ensuring independent project initialization per agent and proper credential handling.

Benefits: Expands QTest toolkit capabilities for comprehensive test management workflows, improves reliability for complex multi-agent scenarios, and enables better discoverability through indexing integration.

Toolkits: Data Analysis Internal Tool¶

Access Pandas data processing capabilities directly as an internal tool without requiring separate toolkit configuration.

What you can do now:

Enable Data Analysis tool: Access Pandas functionality through the "Enable internal tools" dropdown menu (gear icon in chat input), positioned between TODO and Python sandbox tools.
Automatic function activation: Both "Process query" and "Save dataframe" functions are automatically activated when you enable the Data Analysis tool, eliminating manual function selection.
Session persistence: Your enable/disable choice persists throughout the conversation session, maintaining consistent data processing capabilities.
Streamlined workflow: Use Pandas operations directly in chat without navigating to the Toolkits menu or attaching separate toolkit instances.
Consistent experience: Follows the same interaction pattern as other internal tools (TODO, Python sandbox) for familiar, unified usage.

Benefits: Eliminates toolkit creation and configuration overhead for common data analysis tasks, provides instant access to Pandas capabilities in any conversation, reduces friction for ad-hoc data processing workflows, and maintains cleaner toolkit management by moving frequently-used functionality to internal tools.

For more information about this feature, see Data Analysis.

Standardized Field Character Limits¶

Field character limits are now standardized across all entity types with real-time character counters for better input management.

What changed:

Name field limit: All entity Name fields (Agents, Pipelines, Toolkits, MCP Servers, Credentials) now support up to 32 characters.
Description field limit: Description fields across all entities support up to 768 characters.
Message field limits: Conversation Starter and Welcome Message fields support up to 768 characters.
Character counter: Real-time character counter displays remaining characters as you type in any limited field.

Benefits: Ensures consistent field validation across the platform, prevents truncation issues, provides clear visual feedback during input, and helps users stay within limits before saving.

Jira Toolkit: Improved API Version Configuration¶

Jira API version selection is now simpler with dropdown controls, intelligent defaults, and auto-detection during connection testing.

What changed:

Dropdown selection: API version changed from text input to dropdown with v2 and v3 options only.
v3 default: API v3 is now the default (most common for Jira Cloud instances).
Auto-detection: Connection test attempts to detect correct API version and suggests it to users.
Manual override: Users can override auto-detected version when needed.

Benefits: Eliminates API version confusion, reduces setup errors, and streamlines Jira toolkit configuration with intelligent defaults.

Deprecated Features and UI Simplifications¶

Several features and UI elements have been removed or temporarily disabled to simplify workflows and align with the platform's conversation-first direction.

What changed:

RUN tab removed: RUN tab removed from Agent and Pipeline detail views; Configuration tab now opens by default with execution available through Conversations menu.
Monitoring tab removed: Monitoring tab removed from Agent and Pipeline details, and Monitoring sub-menu will be removed from Settings in future release; redesigned project-level monitoring system planned in future.
Pipeline publishing removed: Publishing capability removed for Pipelines (agents-only publication model); pipelines remain private within projects with all creation and execution functionality preserved.
Import temporarily disabled: Import functionality for agents and pipelines temporarily disabled pending improvements.
Profile section redesigned: User Profile menu item removed from sidebar; legacy sections (About Me, Rewards, Own Items, Shared Items) replaced with entity statistics tabs showing created/published counts for Agents, Pipelines, Toolkits, Credentials, MCPs, and Apps.

Benefits: Simplified user interface with clearer execution path through Conversations, reduced feature complexity, focused publication model on Agents only, and aligned with conversation-first platform strategy.

Fixed Issues¶

#3092 Code Node Misformats Code Stored in State Variables: Code node fails to preserve new lines and escaped characters when code is stored as a string in state variables.
#3070 Bitbucket Update File Fails on Full File Rewrite: Full file content replacement can fail with "old content was not found" even when the provided old content matches the file exactly.
#3069 Bitbucket Update/Edit File Fail on JSON Files: JSON file updates fail because file content is read as a parsed dictionary instead of raw text, breaking string-based edits.
#3040 ADO Repos Update File Corrupts Content: Updating files with OLD/NEW markers can duplicate lines and remove original content.
#3005 GitHub Get Commits Diff Tool Throws Exception: GitHub toolkit get_commits_diff throws an exception instead of presenting commit diffs.
#2990 Repository Update File Fails for Multi-line Replacements: Repository toolkits can fail to match and replace multi-line content blocks, preventing updates from being applied.
#2966 Router Node Routes to END Instead of Default Output: Router node incorrectly prioritizes END node over Default Output node when END is included in routes, causing premature pipeline termination.
#2923 ADO Repo Indexing Ignores TypeScript Files in Subdirectories: ADO Repository indexing with "*.ts" whitelist pattern only indexes root-level TypeScript files; files in subdirectories and those with compound extensions (e.g., .steps.ts) are completely ignored.
#2886 GitLab Update File Tool Fails: GitLab toolkit Update File tool fails with "404 File Not Found" error or AttributeError about missing 'extract_old_new_pairs' attribute.
#2871 Figma Toolkit Default Limit Too Low: Figma toolkit default limit parameter (10,000) stops execution prematurely before retrieving complete data from Figma files; should be increased to 1,000,000+.
#2858 Confluence Search Generates Dead Links: Confluence toolkit search results generate links missing "/wiki" path segment, resulting in 404 errors when clicked.
#2799 TestRail Index Search Returns No Results: Search index returns zero results for TestRail toolkit even with cut-off score set to 0.1.
#2782 No Loading Indicator When Uploading Attachments: File attachments upload without visual feedback (spinner, progress bar, or status message), leaving users unable to tell if upload is in progress, stuck, or complete.
#2780 Stepback Summary Index Fails with Anthropic Models: Stepback Summary Index fails with Anthropic models due to duplicate tool call IDs on retry.
#2763 OpenAPI Default Parameters Ignored: OpenAPI toolkit ignores default parameter values when making requests.
#2715 Code Node Fails with Inline Newline Characters: Code node execution fails when code contains inline \n characters.
#2693 Nested Agent Tool Call Fails: Parent agent calling nested agent with toolkit fails with "This event loop is already running" error.
#2671 ID Conflicts Between Public and Custom Agents/Pipelines: Public agents can share IDs with custom agents/pipelines, causing conflicts when adding both to conversations.
#2670 Bitbucket Repository Field Character Limit: Bitbucket toolkit Repository field limited to 36 characters.
#2659 ServiceNow Toolkit Fails with Valid Credentials: ServiceNow toolkit produces errors even with valid credentials configured.
#2639 Router Node Error When Routing to END: Router node produces errors when first routing branch points directly to END node.
#2621 QTest Project ID Not Marked Required: QTest toolkit Project ID field not marked as required despite being mandatory.
#2619 Validation Error with Disabled Messages State: Pipeline produces validation error when messages state is disabled.
#2578 Pipeline Flow Loops Back After Interrupt: Pipeline flow incorrectly returns to starting node after interrupt instead of continuing forward.
#2571 No Full Screen View for Index Tool Configuration Fields: Index tool configuration fields (Filter, Chunking Config, etc.) lack full screen editing capability, making it impossible to properly review and edit complex JSON configurations in small textarea boxes.
#2569 Default Cut-off Score Too High for Search Indexes: Search index tools (Search Index, Stepback Summary, Stepback Search) use default cut-off score of 0.5 instead of recommended 0.2, potentially filtering out relevant results.
#2563 Router Node Ignores Conditions: Router node condition logic ignored, always routes to specific node regardless of condition evaluation.
#2540 Invalid Tools Not Highlighted in Red: Toolkit interface does not highlight non-existing or invalid tools in red.
#2532 GitLab Index Creation Fails: GitLab toolkit cannot create or reindex indexes with "AttributeError: 'NoneType' object has no attribute 'get_index_meta'" error.
#2525 SQL Toolkit Save Button Unresponsive Without Dialect: Save button in SQL toolkit creation does not work when Dialect field is left unselected, with no error message indicating the field is required.
#2518 CodeMirror Search Dialog Disappears: CodeMirror search dialog (CMD+F) appears briefly and immediately disappears, making search unusable.
#2512 Pipeline Agent Node Rendering Issues: Agent node in pipelines displays incorrectly with rendering problems.
#2504 Credential ID Not Updating: Credential ID field does not update when credential name is overwritten.
#2503 ADO Wiki Update Response Not Handled: Azure DevOps toolkit does not properly handle API response during Wiki page updates.
#2498 State Variable Values Not Applied with Legacy Type Definitions: Pipeline state variable values not applied when YAML uses legacy type definition "string" instead of current "str" from version 1.6.
#2492 Canvas Attachment Binding Not Persisted: Agent canvas shows "attachment manager set successfully" when binding artifact toolkit, but binding disappears after reopening canvas; requires page refresh to display correctly.
#2496 State Dropdowns Unsorted: Custom state dropdowns appear in random order with default states (input/messages) not pinned to top.
#2444 Modified Secret Received Unmasked: Modified secrets received unmasked from backend, exposing sensitive values.
#2418 LLM Node Ignores Step Limit: LLM node hard-codes 15 execution iterations, ignoring configured Step limit in Agents and Pipelines.
#2417 Project Admin Sees Only Own Executions: Project administrators can only view their own execution history instead of all project executions.
#2415 Default State Variables Become Inactive: Input and Messages default state variables incorrectly become inactive when creating new custom state.
#2407 LLM Model Breaks When Clearing Token Fields: Clearing "Context Window" or "Max Output Tokens" fields in LLM model configuration causes 500 error and model becomes unavailable; workaround is to delete and recreate the model with proper values.
#2405 No Upload Progress for Large File Attachments: No upload progress indicator shown when attaching large files (5MB+) in conversations, causing UI to appear frozen; files eventually upload successfully.
#2375 Multi-File Attachment Rejected with Single Duplicate: When selecting multiple files for chat attachment, all files (including unique ones) are rejected if one duplicate is present; workaround is to deselect duplicate files before attaching.
#2366 Bucket Selection Changes Unexpectedly: Bucket selection changes when clicking dropdown options in artifact configuration.
#2363 Cannot Read Python Files from SharePoint: SharePoint toolkit unable to read .py files from SharePoint libraries.
#2360 Deprecated Prompt Field in New LLM Nodes: Newly created LLM nodes include deprecated "prompt" field in YAML alongside new "system" and "task" fields; deprecated field should be ignored.
#2352 SharePoint Toolkit Read List Fails with Graph API: Read List functionality fails with error when SharePoint toolkit is configured to use Graph API; workaround unavailable.
#2351 Zephyr Essentials Reindexing Removes Previous Data: Reindexing in Zephyr Essentials toolkit deletes previously indexed documents even when "Clean Index" is not selected; existing data is removed and reindexed.
#2350 Misleading Error Message for Token Limit: Agent shows "Maximum tool execution iterations (15) reached" when actual issue is Max Output Tokens limit; increase Max Output Tokens in Anthropic model settings to resolve.
#2347 Pandas Bucket Name with Dash Causes Error: Pandas toolkit fails when bucket name contains dash character.
#2294 Misleading Unsaved Changes Warning: Starting conversation with Canvas open triggers unsaved changes warning even when no edits were made.
#2053 Claude Model Integration Error with Dial Key: Claude model integration via Dial API key fails with "invalid structure" error on stream_options path.
#2051 Gemini/Claude Models Fail via Dial API: Integrating Gemini or Claude models using Dial API key results in "invalid structure" error on stream_options path.
#1986 ADO Attachment Reading Not Available: Azure DevOps toolkit lacks capability to read attached images from work items.
#1869 Zephyr Essential Tools Return 400 Error: Multiple tools in Zephyr Essential toolkit fail with "400 Client Error: Bad Request" including Create/Update operations for test cases, cycles, and executions.
#1746 Jira Attachment Content Limited to JSON/TXT: Get Attachments Content tool in Jira toolkit does not support common file types (PNG, JPEG, XLSX); only JSON and TXT files are supported.
#1547 Jira JQL Search Fails Without Priority Field: Jira toolkit fails with validation error when searching issues that lack the "priority" field; affects Jira configurations where priority field is disabled.
#1531 Multiple Agents with Same Toolkit Type Interfere: Multiple agents using same toolkit type (ADO Boards, GitLab, QTest) in same pipeline/agent interfere with project initialization.
#1371 OpenAPI Toolkit Fails in Pipelines: OpenAPI Toolkit cannot be used in pipeline function or tool nodes; execution fails with errors.

Known Issues¶

#2389 Browser Toolkit Fails Without Credentials: Browser toolkit execution fails when saved without credentials selected, even when Google tool is not selected; workaround is to add credentials.
#2304 MCP Client Disconnects on macOS: MCP client disconnects from platform despite tray showing "connected" on macOS; logs show "packet queue is empty" errors and repeated disconnect/reconnect cycles.
#1990 Browser Toolkit Fails Without Google Credentials: Browser toolkit execution fails with "NoneType object is not a mapping" error when Google tool is not selected and credentials are missing.
#1163 MCP Only Executes Latest Version: MCP-tagged resources execute the latest version regardless of which version has the MCP tag; only the latest version is accessible via MCP.