LangChain revamps human-in-the-loop middleware with typed schemas for tool interrupts
Hacubu · x · 2026-10-16
LangChain's Sydney Runkle announced a revamped human-in-the-loop middleware: interrupts on tool calls now ship with typed schemas, so human responses are validated before returning to the agent, and schemas can render nicer approval forms in your UI.
The docs define four decision types — approve, edit, reject, and respond (a synthetic tool result for ask-user tools). Middleware checks each tool call against a configurable interrupton policy, pauses execution when needed, and uses LangGraph's persistence layer to safely save state and resume later.
More from coding & agent
- Moving agent work from GPUs to CPUs for durable, predictable actions — cramforce · 2026-10-16
- Jev tutorial: 106x cheaper, 180ms yes/no gates replace PASS/FAIL string matching — max_paperclips · 2026-10-16
- Stop giving agents raw API keys: prepaid balances and per-credential caps — Cold_Hall_5384 · 2026-10-16
- Six-month prediction: hiring for generic agent harnesses built from scratch will cool — yi111 · 2026-10-16
- Omarchy ISO wouldn't boot on NVIDIA, so he had Claude patch it — indragie · 2026-10-16
- Dev uses Grok to clean up 3D printer firmware, catches it sneaking in nonsense code — yacineMTB · 2026-10-16