feat(proxy): add connection pooling, HTTP/2, and multi-worker support

feat: add streaming memory tool support with credential error handling

fix: correct cost calculation for cache tokens and simplify memory DevEx

fix: resolve mypy type errors across codebase

Add hierarchical memory system with graph + vector storage

fix: parse actual token usage from SSE streams instead of byte estimation

Add missing pillow dependency for image compression

Add image token compression with trained ML router

fix: account for cached tokens in cost calculation across all providers

Add E2E tests for Google multimodal content preservation

Fix Google Gemini conversion to preserve non-text content (Gap #4)

Fix mypy type checking errors in CCR batch modules

Update compressor CCR markers and add batch compression eval

Add multi-provider batch API support with CCR post-processing

Remove hardcoded source hint system from ContentRouter

Add hnswlib to dev dependencies for CI testing

Make hnswlib optional and skip tests when unavailable

Update memory documentation for HierarchicalMemory system

Replace legacy memory system with HierarchicalMemory

Add comprehensive evaluation framework for compression accuracy

Add pre-commit hooks for ruff linting and formatting

Update test for lowered TOIN confidence threshold default

Fix TOIN integration and add persistence support

Add reasoning capability forwarding to HeadroomAgnoModel

Add dynamic anchor selection with content deduplication

Add pluggable storage backend abstraction for CompressionStore

Add demo GIF to README

Bump version to 0.2.15

Fix streaming responses not tracking costs

Add architecture diagrams to README and docs

Add PyPI downloads badge to README

Integrate DynamicContentDetector into CacheAligner (Phase 1)

Fix cost tracking to use LiteLLM pricing database correctly

Add demo video link to README

Add quality retention eval and fix linting for Python 3.12

Bump version to 0.2.14

Merge pull request #8 from nicolabeghin/proxy-custom-openai-api-url

Merge pull request #9 from nicolabeghin/docker-support

Merge pull request #6 from smartwatermelon/claude/docs-macos-deployment-20260119

feat: add Docker configuration for containerized deployment

docs update

feat(config): add OpenAI API URL override option

fix(deployment): correct port placeholder in LaunchAgent plist template

docs(deployment): add macOS LaunchAgent deployment guide and templates

Add Phase 2 Progressive Summarization and fix Agno integration tests

Fix HeadroomAgnoModel to optimize tool outputs at invoke level

Rewrite README with verified before/after examples

Implement COMPRESS_FIRST strategy for IntelligentContextManager

Add IntelligentContextManager for semantic-aware context management

Bump version to 0.2.13