← Code Exchange
Zeitlich - An opinionated AI agent implementation on top of Temporal
Zeitlich is a lightweight framework for building reliable, stateful AI agents. It provides the building blocks for creating multi-turn agents. Think OpenClaw, but durable, with long-term state, and enterprise-ready.
Building production AI agents is hard. Agents need to:
- Survive failures — What happens when your agent crashes mid-task?
- Handle long-running work — Some tasks take hours or days
- Coordinate — Multiple agents often need to work together
- Maintain state — Conversation history, tool results, workflow state
Temporal solves these problems for workflows. Zeitlich brings these guarantees to AI agents.
Features#
- Durable execution — Agent state survives process restarts and failures
- Thread management — Redis-backed conversation storage with automatic persistence
- Type-safe tools — Define tools with Zod schemas, get full TypeScript inference
- Lifecycle hooks — Pre/post tool execution, session start/end
- Subagent support — Spawn child agents as Temporal child workflows
- Filesystem utilities — In-memory or custom providers for file operations
- Model flexibility — Framework-agnostic model invocation with adapters for LangChain (and more coming)
