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

Fix agno model.py mypy errors by declaring name and provider fields

Fix test_on_llm_error to pass required run_id argument

Fix mypy type errors for Python 3.12 compatibility

Fix HeadroomAgnoModel to extend agno.models.base.Model and add real integration tests

Enhance /stats endpoint and add comprehensive benchmarks

Fix ruff formatting

Add TOIN field-level learning and comprehensive integration tests

Bump version to 0.2.10

Fix mypy type errors in LiteLLM and OpenAI providers

Bump version to 0.2.9

Add feature coverage and known limitations to Agno docs

Add Agno integration documentation

Add Agno integration tests to CI workflow

Add Agno agent framework integration

Use LiteLLM for model pricing instead of hardcoded values

Fix proxy passthrough routing to use SDK headers instead of User-Agent

Bump version to 0.2.8

Fix mypy type annotation error in content_router.py

Bump version to 0.2.7

Add CCR marker support to all compressors and fix retrieval

Fix LLMLingua API: use context=[content] not original_prompt

Fix 'list' object has no attribute 'strip' error

Bump version to 0.2.6

Enable LLMLingua and code-aware compression by default in proxy

Bump version to 0.2.5