All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

• Add LangChain runtime adapter with chain support (invoke/run) and output normalization.
• Add session-based chat support; split respond(session) and next(state) APIs.
• Unify adapter interface for both Run and Chat via RuntimeAdapter.
• Support custom adapters via agent.adapter = "module:attr"; enforce explicit inheritance from RuntimeAdapter.
• Wire custom adapters for both run and chat; add example examples/custom_adapter and integration test.
• Add GitHub Actions CI for tests, Ruff, mypy (via pre-commit) and PyPI publish workflow.
• Fix Ruff issues and tidy imports/examples; minor README updates.
• CLI: add uai run cancel and uai run stop to cancel runs.
• API: allow adding artifacts to runs (POST /run/{id}/artifacts) and chat sessions (POST /chat/{session}/artifacts).
• Utils: expose post_log, add_run_artifact, add_chat_artifact, and request_human_input to let user code manage logs/artifacts and input waiting.

• Initial release with FastAPI service, run queue (inline/Procrastinate), and CrewAI + callable runtimes.
• Examples for callable, CrewAI, and human-in-the-loop CrewAI.