One Decorator Away From Production-Ready AI Agents

Short summary

ToolOps is a Python package that adds production infrastructure to AI agents via decorators—caching, semantic deduplication, request coalescing, and observability. Supports vector-based semantic caching to reduce LLM calls by up to 90%, with PostgreSQL and OpenTelemetry for distributed systems. Open-source on GitHub with two decorators covering all use cases.

• •Two decorators (@readonly, @sideeffect) wrap agent code with production infrastructure without code changes
• •Semantic caching using vectors collapses similar queries, reducing LLM calls up to 90%
• •Supports in-memory, file, or PostgreSQL backends with OpenTelemetry observability

Generated with AI, which can make mistakes.