syrin-labs/syrin-python

Based on the README, here's the technical summary: --- Built on a cost-aware architecture with pre-call token estimation and post-call recording from providers, supporting multi-tier rate limiting (per-run, hourly, daily, monthly) and shared budgets across parallel agents. Provides pluggable memory backends (SQLite, Qdrant, Redis, PostgreSQL) with semantic search, and 72+ lifecycle hooks for fine-grained observability into LLM calls, tool execution, and budget thresholds. Integrates with OpenAI and Anthropic models, with orchestration patterns including agent handoffs, spawning, and dynamic pipelines that let the LLM decide agent routing.

Available on PyPI.