gotoHuman

gotoHuman is a hosted human-in-the-loop platform that lets AI agents pause and request a human review before taking consequential actions. Developers define review templates, the agent calls the API when it hits a decision point, and a reviewer approves, edits, or rejects the output from an inbox, email digest, or Slack notification. The result is posted back to the workflow via webhook so the agent can continue. It supports multi-step approvals, artifact versioning to compare AI outputs, manual triggers from custom forms, and agent memory that feeds prior reviewer feedback into future runs.

It's aimed squarely at teams shipping autonomous agents who can't yet let them run unsupervised, whether they're built in no-code tools like n8n and Make.com or in code via the JavaScript and Python SDKs, LangGraph, or Anthropic's MCP. There's a free tier with paid plans listed on the pricing page, and the infrastructure leans enterprise-friendly: European servers, GDPR compliance, encryption at rest and in transit, and SOC 3 plus ISO 27001 certifications. It is not a model or an agent framework itself; it is the approval queue you bolt onto whatever stack you already have.