[Release] crewaiInc/crewAI 1.14.7: 1.14.7

crewAI v1.14.7 introduces pluggable backends for memory, knowledge, RAG, and flow management, along with a native Snowflake Cortex LLM provider and chat API for conversational flows. The release includes significant bug fixes for checkpoint restoration, telemetry, file input handling, and security vulnerabilities. Performance improvements focus on lazy-loading docling imports to speed up crewai imports, while refactoring decouples conversational logic and splits flow.py into modular components.

Key Points

• •Pluggable default backends enable flexible configuration of memory, knowledge, RAG, and flow systems
• •New Snowflake Cortex native LLM provider and Snowflake integration guide expand enterprise database support
• •Chat API for conversational flows enables interactive, multi-turn dialogue capabilities
• •Type DSL triggers as route-aware decorators improve flow control and type safety
• •Checkpoint restoration fixed to properly rebuild custom BaseLLM instances and prevent unintended replay
• •Runtime state scoped per run prevents memory growth and isolates concurrent executions
• •Security: Resolved pip-audit CVEs in aiohttp, docling, and docling-core dependencies
• •Performance: Lazy-loading of docling imports significantly improves crewai import speed
• •Refactored flow.py into separate DSL, definition, and runtime modules for better maintainability
• •Enhanced LLM events now surface real finish_reason, sampling parameters, and response.id for better observability

Workflow Diagram

Concepts