Releases: presidio-oss/specif-ai
Release v2.8.1
2.8.1
✨ Enhancements
-
Clickable Requirement Badges
Requirement badges now support direct clicks to instantly open linked Jira or ADO tickets — making navigation faster and more intuitive (#378). -
Updated Documentation
The Docusaurus home page and Core Features section have been refreshed for improved structure and clarity (📚 Explore the enhanced documentation →) (#369).
🛠 Fixes
-
ID Duplication & Pagination Issues
Resolved duplicate ID generation and pagination inconsistencies in PMO integrations for more reliable data handling (#377). -
Chat Streaming & Anthropic Model Support
Fixed an issue that prevented native Anthropic Claude 4 and Claude 3.7 models from functioning correctly during chat streaming (#365).
v2.8.0
2.8.0
Added
-
Introduced Strategic Initiatives: Craft compelling business proposals that connect high-level strategy to actionable deliverables. (#317)
📘 Explore how to draft proposals with SI → -
Added AI-Powered Inline Editing (Beta): Instantly enhance any text within editable fields using AI assistance—without disrupting the workflow. (#317)
⚡ Supercharge your edits with AI-powered inline editing →
Enhancements
-
Jira integration now mirrors the ADO workflow, enabling precise Push/Pull actions at the ticket level directly from the PRD list for a smoother, more consistent experience. (#343)
🔄 Learn how Jira actions are now more seamless → -
Enhanced HAI Chat to fully replace requirement text with structured AI suggestions, enabling clearer edits and reducing hallucinations during refinement. (#354)
💬 Discover smarter refinement with HAI Chat → -
JIRA and ADO credentials are now stored securely in the Electron app, improving security and eliminating reliance on metadata. (#345)
Fixed
v2.7.0
2.7.0
Added
- Bi-directional sync between JIRA and Specifai ensures consistent updates across platforms. (#284)
- Azure DevOps (ADO) integration enables seamless pull and push operations. (#313)
- Legacy JIRA IDs in requirement files are now automatically migrated to a unified PMO ID format for cross-platform support. (#326)
- Test cases can now be generated using an Agentic approach based on the User Story. (#312)
- User Guide documentation link is available in the About section for easier onboarding and reference. (#325 )
- Solution creation now automatically redirects to the solution view with real-time progress updates, along with retry and cancel options. (#252, #280)
Enhancements
- The rich text editor for requirement descriptions now supports full table and link management. (#284)
- AI-generated content now retains formatting, including tables and links. (#303)
- Navigation between PRDs and related user stories is now smoother and more intuitive. (#315)
- Export options are grouped for a more streamlined and efficient experience. (#284)
- Added Reset button to the settings page to restore AI and configuration defaults. (#314)
Fixes
- JIRA callback port set to 12345 to avoid conflicts with any existing system processes. (#334)
- Resolved an issue with saving LLM configuration when upgrading from older 1.x versions. (#314)
- AI chat header now consistently displays “Talk to HAI” as part of updated branding. (#243)
- Card layouts now maintain consistent height across pages for improved visual alignment. (#267)
v2.6.0
v2.6.0
Added
- Introduced a visual view of the AI’s thinking process during story and task generation/regeneration for better transparency. (#223)
- Added support for two advanced AI models — Claude 4 Sonnet and Opus — for improved outcomes. (#239)
Enhancements
- Enhanced traceability by displaying detailed thinking steps from the AI. (#238)
- Linked AI thinking with visible UI actions, making progress easier to understand. (#223)
- Improved user experience by allowing the app header, including the login screen, to be dragged and repositioned. (#227)
Fixes
- Refined dropdowns in LLM settings for a cleaner and more user-friendly selection experience. (#124)
v2.5.0
v2.4.0
v2.4.0
Release Notes
Added
- Upgraded the AI chat experience to be more interactive and responsive, including MCP tool integration and real-time streaming capabilities. (#185)
- Added agent-based workflow for better chat suggestions. (#180)
- Integrated HAI-Guardrails for enhanced security. (#209)
Enhancements
v2.3.0
v2.3.0
Release Notes
Added
- Added support for configuring Model Context Protocol (MCP) servers in the Solution's Integrations section: (#168)
- Used in both Create Solution and Generate Stories & Tasks workflows
- Includes predefined support for AWS Bedrock
- Supports adding multiple custom MCP servers
- Implemented AI-Agentic workflow to generate tasks and user stories. (#177)
Enhancements
- Switched to minimum threshold for preferred count in solution requirement generation. (#194)
- Enhanced the Home Screen UI for improved visual experience when no solutions are present. (#193)
Fixes
- Resolved cross-region model connection issues with AWS Bedrock. (#182)
- Updated the login process to require users to select a working directory, ensuring proper setup before proceeding. (#194)
- Fixed scrolling issues on the Solution Integrations page. (#194)
- Fixed Analytics Settings toggle to capture user input, improving observability tracking across the platform. (#189)
v2.2.2
v2.2.2
Release Notes
Fixes
- AI Chat integration was removed from the "Add" requirement document flow.
- Enhanced Business Process navigation with improved guards to prevent data loss and eliminate redundant popups.
v2.2.1
v2.2.1
Release Notes
Fixed
- Resolved an issue with the Create Solution when using AWS Bedrock models. Replaced the community package with the official @langchain/aws package and correctly configured the maxTokens parameter within the expected input path for the Bedrock model.
v2.2.0
v2.2.0
Release Notes
Added
- Implemented Agentic flow within the Create Solution workflow utilising LangGraph. (#128)
Enhancements
- Disabled search functionality for the provider dropdown to streamline user interaction. (#160)
- A clear button was introduced to the select component for improved usability. (#146)
Fixed
- Addressed an issue causing tasks to regenerate upon cancellation. (#136)
- Standardised font consistency in the Export dropdown menu by transitioning from Angular Material to Tailwind CSS. (#136)
- Resolved a click-related issue in the Export dropdown menu. (#159)
- Eliminated an unintended discard pop-up during the Add and Edit Requirement workflows. (#161)