Python SDK for Cycles — reserve, commit, release budget and risk for AI agents
Execution safety & governance layer for AI agents: idempotency, budgets, tool control, audit logs.
Production-grade guardrails for AI agent function calls. Prevent overspending, unauthorized purchases, unsafe deletions, and more — with Human-in-the-Loop approval.
Small reliability layer for HTTP APIs and LLM calls. Idempotent HTTP/LLM proxy with retries, cache, circuit breaker and predictable AI costs.
Adaptive inference budget controller for self-hosted LLMs. Controls thinking tokens, tracks GPU cost per query.
