v0.9.1

• SSE transport support for remote MCP access
• Fixed AI Judge Score Threshold Not Applied
• Fixed history persistence bug
• License updated to ELv2

v0.9.0 (Swordfish)

• CLI tooling - run tests, manage test cases from terminal
• MCP server - integrate with Claude Code and AI assistants
• Test case CRUD API endpoints
• Token auth support for projects endpoint
• Improved auth security - validate before DB connection

v0.8.0 (Blobfish)

• Grok (xAI) provider support - Grok 3, Grok 2 and variants
• DeepSeek provider support - DeepSeek Chat and DeepSeek Reasoner
• Run single test case - click Play button on any test case row to run it individually
• Run selected test cases - use checkboxes to select multiple test cases and run them together
• Drag-to-reorder test cases - drag the grip handle to reorder test cases in a suite
• Suspend/resume test cases - toggle individual test cases to skip them during runs without deleting
• Fixed issue with ignoring score threshold
• LLM Judge now supports all providers (Grok, DeepSeek added)
• UI improvements
• JSON syntax highlight

v0.7.1

• Prompt preview in test suite
• "Latest" prompt version option for test suites
• Change prompt version in test suite settings
• "Is JSON" and "Contains JSON" validation rules
• Documentation: Added ENCRYPTION_KEY generation instructions to README and Dockerfile
• Improved error messages for network failures (now shows root cause like DNS/connection errors)
• Run tests multiple times (N iterations) for consistency testing
• UI improvements

v0.7.0 (Mahi-mahi)

• Open Source now.