v0.2.2

Added

• Latency Tracking: Track execution duration for each tool call
  • @stat.track decorator for automatic latency tracking (recommended)
  • async with stat.tracking(name, type) context manager alternative
  • duration_ms parameter in record() for manual timing
  • New latency columns: total_duration_ms, min_duration_ms, max_duration_ms
  • latency_summary in get_stats() response with total duration
  • Per-tool avg_latency_ms, min_duration_ms, max_duration_ms metrics

Changed

• Database schema bumped to v3 with latency tracking columns
• get_stats() response now includes latency_summary object
• Each stat item now includes latency fields
• API Improvement: @stat.track decorator is now the recommended way to track calls
  • Eliminates the "first line" requirement
  • Automatic latency measurement
  • Never fails user code

Migration

• Automatic database migration from v2 to v3
• Preserves all existing data
• New latency columns default to 0/NULL for existing records

v0.2.1

Added

• Token Tracking: New feature to track response sizes and estimate token usage
  • response_chars parameter in record() for automatic token estimation
  • input_tokens and output_tokens parameters for actual token tracking
  • New report_tokens() method for deferred token reporting
  • Token summary in get_stats() response with totals and averages
• Schema Migration: Automatic database migration from v1 to v2
  • Adds token tracking columns with backward compatibility
  • Preserves all existing data
• Documentation: Added new and updated guides, API references, and examples

Changed

• Database schema bumped to v2 with new columns:
  • total_input_tokens, total_output_tokens
  • total_response_chars, estimated_tokens
• get_stats() response now includes token_summary object
• Each stat item now includes token fields and avg_tokens_per_call

v0.1.2

Added

• GitHub community files: CONTRIBUTING.md, SECURITY.md, issue/PR templates
• Tags System documentation with filtering examples
• Expanded API Reference with complete code examples
• Test coverage improved

Changed

• Improved README.md structure and navigation with collapsible MCP client configs
• Enhanced llms.txt for better AI agent consumption

v0.1.1

Bug fix release.

Fixed

• Python 3.10 compatibility: replaced datetime.UTC with datetime.timezone.utc

v0.1.0

Initial release.

Added

• Core: MCPStat class - unified API for usage tracking and analytics
• Database: MCPStatDatabase - SQLite-backed storage with async-safe operations
• Logging: MCPStatLogger - optional file-based audit logging (pipe-delimited format)
• Built-in tools for exposing stats to MCP clients:
  • get_tool_usage_stats: Query tool usage statistics with filtering
  • get_tool_catalog: Browse registered tools with tags, search, and metadata
• Prompts: generate_stats_prompt() for LLM-friendly usage reports
• Helpers: build_tool_definitions() and build_prompt_definition() for easy MCP registration
• Handler: BuiltinToolsHandler class for handling built-in tool calls
• Utils: normalize_tags() with stopword filtering, derive_short_description() utilities
• Metadata enrichment system with tags and descriptions
• sync_tools(), sync_prompts(), sync_resources() for automatic metadata sync from MCP objects
• type_filter parameter in get_tool_usage_stats for filtering by primitive type
• Environment variable configuration support (MCPSTAT_DB_PATH, MCPSTAT_LOG_PATH, MCPSTAT_LOG_ENABLED)
• Async-first design with thread-safe internals
• No required dependencies (pure Python stdlib)
• Optional mcp extra for MCP SDK integration
• Full type annotations with strict mypy compliance (py.typed marker included)
• Comprehensive test suite