v1.2.0

• Repository Cloning & Exploration: The AI now clones your repository and generates a directory listing to understand the full codebase structure
• read_file Tool: AI can read files with line ranges to understand context, verify imports, and trace function calls
• grep_search Tool: AI can search for patterns across the codebase using grep, enabling comprehensive code analysis
• Multi-Provider Support: Advanced reviews work with both Google Gemini and Anthropic Claude providers
• Configurable Iterations: Control how many tool calls the AI can make during analysis

• Deeper, more contextual code reviews
• AI understands how changes affect other parts of your codebase
• Catches cross-file issues and dependency problems

/matter fix

1. Collects all unresolved review comments on the PR
2. Reads the current file contents for affected files
3. Uses AI to generate targeted code fixes
4. Creates a new PR with the fix branch automatically

• Claude Sonnet 4.5 (preferred)
• Gemini 3 Pro Preview
• Axon Code (MatterAI)

• File-Based Memory Matching: Memories are matched based on files changed in the PR
• Usage Tracking: Automatically tracks which memories are used and how often
• PR Summary Display: Shows which repository memories influenced the review
• Prompt Injection: Relevant memories are injected into AI prompts for contextual reviews

• usage_count: Track how often each memory is used
• last_used_at: See when memories were last referenced

### 📚 Repository Memories Used

**1. Authentication Guidelines**
   - *Tags:* security, auth
   - *Files:* src/auth/*.ts
   - *Content:* Always validate JWT tokens...

• Repository Filter — Filter logs by specific repositories or by MCP Server requests
• Provider Filter — Filter by LLM provider (OpenAI, Anthropic, Google/Vertex, MatterAI)
• Model Filter — Filter logs by specific AI models
• Prompt ID Filter — Filter by prompt identifiers for tracing specific request types

• OpenAI
• Anthropic
• Google / Vertex AI
• MatterAI

• New Advanced Reviews settings section in the dashboard
• Multi-repository selection via autocomplete
• Save/update controls with loading states
• Fetches and displays currently enabled advanced review repositories

• Track estimated developer time saved through automated reviews
• Integration with dashboard overview metrics
• Helps justify ROI for enterprise AI tooling investments

• Product Showcase Cards — Visual cards highlighting key products:
  • AI Code Reviews
  • Axon Code AI Agent
  • Model Inference Gateway
• New Hero Background — Modern, premium visual design
• Version Display — Application version now shown during signup
• Enhanced Footer — Updated branding and product information

• Updated base color palette for improved contrast and readability
• Refined card component styling
• Enhanced input and select component borders
• Improved sidenav visual hierarchy
• Consistent gradient border styling across components
• Updated dashboard navbar styling

• Updated statistics cards with new iconography
• Improved date range filter presentation
• Enhanced PR data visualization
• Cleaner table layouts across all views

• Cost by Provider chart respects all active filters
• Requests by Model chart respects all active filters
• Daily Cost Trend reflects filtered data
• Daily Request Volume reflects filtered data

• Updated to Gemini 3 Pro Preview for advanced reviews
• Updated to Gemini 3 Flash Preview for standard reviews
• Improved model routing based on task complexity

• Fixed accurate token counting for all providers
• Improved token storage for billing and analytics
• Corrected input/output token separation

• Updated multiple core dependencies for security and performance
• React ecosystem packages updated to latest stable versions
• Added @google/genai - Google Generative AI SDK for advanced Gemini integration
• Build configuration optimizations

• Optimized LLM logs data filtering with memoized computations
• Improved chart rendering performance in metrics views
• Reduced re-renders in settings panels

• Fixed filter state persistence across tab navigation
• Resolved chart rendering issues with empty data sets
• Fixed select component dropdown arrow alignment
• Corrected toast notification z-index layering
• Standardized check run names to “Matter AI Code Review”
• Removed duplicate check runs issue

​

Feature Flags

​

Audit Logging

​

Multi-tenant Considerations

​

Environment Variables