Changelog

Product Updates & Release Notes

Follow along as we build the AI ecosystem for software delivery — new features, improvements, and platform updates, day by day.

April 20, 2026

WalnutAI 4.3.12

Enhancements

Improved Agent Reliability for Code Editor

  • Background agents now run beyond 120 seconds in a fully detached mode and notify upon completion
  • No forced timeouts - stop tasks anytime
  • Live progress tracking (e.g., iteration status)
  • A new verification subagent validates results before completion
  • Sub-agents now support streaming with instant cancellation

No more hallucinated answers in Code editor

  • The model avoids generating values or results without proper evidence
  • Responses are re-verified when challenged (e.g., “Are you sure?”)
  • Empty responses now display clear errors instead of blank outputs

UI Overhaul for Code Editor

  • Clean, modern interface with a minimal design
  • Collapsible tool activity for improved readability
  • Queue messages while agents run and continue working without interruption
  • Quick actions: / for command palette and ? for keyboard shortcuts
  • Syntax highlighting with expandable code blocks
  • Compact permission popups with clear risk labels
  • Automatically hides temporary agent-generated files

Faster performance

  • Smoother and faster streaming with reduced overhead
  • Handles long conversations (500+ messages) without slowdown
  • Smarter prompt caching reduces token usage
  • Background compaction improves performance without interruptions

Smarter test recording

  • Tab switches are reliably captured during recording
  • “Continue on failure” is consistently respected
  • Every step clearly displays its expected result

More reliable element detection

  • Accurately distinguishes similar elements across different pages
  • Identifies repeated actions (e.g., multi-step “Save” buttons) uniquely
  • Introduces a DOMPath fallback locator for improved detection
  • Enhanced attribute handling for more stable test steps

Smarter AI assistance

  • Supports large dataset generation without truncation issues
  • Smarter failure handling using on-screen error messages

Additional update

  • Walnut Agent Setup 0.4.5.exe included

April 7, 2026

WalnutAI Release

New

Figma Integration

  • Connect your Figma account directly to WalnutAI - no exports or copy-pasting required
  • Select screens and automatically generate user stories and test cases
  • Eliminates manual translation of designs into requirements

Code Repository Integration

  • Connect WalnutAI directly to your code repository
  • Generate user stories and test cases based on actual code
  • Keeps documentation aligned with implementation

Requirements Traceability Matrix (RTM)

  • Automatically builds RTM in the background
  • Tracks epics, features, stories, test cases, and defects
  • Provides full end-to-end traceability

Test Data Generation

  • Automatically generates complete datasets for testing
  • Speeds up testing with tailored data
  • Shortcut: Use @ in AI Healer → select Test Data Generation

Enhancements

Multimodal LLM Integration

  • Support for multiple LLM models within a project
  • Auto Mode selects the best model automatically
  • Improves accuracy across modules

Jira Integration - Redesigned

  • Simplified settings with cleaner UI
  • Bidirectional live sync with Jira
  • Hierarchy support for Epics, Features, Stories
  • Re-sync support and assignee mapping
  • Custom issue types supported

Azure DevOps - Bidirectional Sync

  • Real-time sync of work items and updates
  • Supports attachments and comments
  • Manual sync option available
  • Available as Beta feature

Multi-Platform Testing Support

  • Supports API, Database, Mainframe, Android, Hybrid
  • Use Smart Recording for quick setup
  • Unified testing across platforms

Bug Fixes

Intelligence Hub

  • Fixed file upload and generation performance issues
  • Correct structured output for TXT uploads
  • Improved DOCX, PDF, XLSX, PPT handling
  • Resolved duplicate uploads and chatbot issues

Test Case Execution

  • Fixed stuck execution states
  • Improved IF/ELSE validation and execution flow
  • Prevented duplicate reports

Test Case Management

  • Resolved parameter and mapping issues
  • Fixed duplicate creation bugs
  • Improved bulk actions

General Improvements

  • Restored filters across modules
  • Improved large payload handling
  • Added confirmation dialogs for Git actions