# Humangent — Full Context for AI Tools

> Human-in-the-loop for n8n — managed, not duct-taped.

## What is Humangent?

Humangent is a human-in-the-loop (HITL) platform built specifically for n8n workflow builders. It provides a dedicated, purpose-built layer for managing human decisions in AI agent workflows — with routing, prioritization, escalation, configurable timeouts, custom actions, and a centralized inbox.

When you build an AI agent workflow in n8n that touches real customers, records, or money, you often need a human to review and approve a proposed action before it executes. Humangent is the dedicated layer for that: it treats human decisions as things that need to be managed, not one-off events scattered across workflows.

## How It Works

1. Your n8n workflow reaches a step that requires a human decision.
2. You add a standard HTTP Request node that POSTs the proposed action to Humangent.
3. Your workflow pauses at a Wait for Webhook node.
4. Humangent notifies the designated reviewer via their preferred channel.
5. The reviewer logs in to their Humangent account and opens the request in the centralized inbox (or follows a direct link).
6. They choose an action — approve, reject, escalate, request revision, or any custom action defined in the workflow.
7. Humangent fires a callback to your n8n webhook with the chosen action and any edited values.
8. The Wait node wakes up and the workflow branches based on the decision.

## Key Differentiators

- **Centralized inbox.** One dashboard showing every pending decision across every workflow. Filter by status, reviewer, or deadline. Drill into any request to see its full history.
- **Managed, not fire-and-forget.** Assign each decision to a specific person, team, or role. Set a deadline. If nobody acts, Humangent escalates to a secondary reviewer automatically.
- **Custom actions.** Not just approve/reject. Define any set of actions your workflow needs — approve, reject, escalate, request revision, or anything else. Each routes to a different branch in n8n.
- **Simple reviewer experience.** Reviewers log into Humangent directly — no n8n account, no Slack seat. They see the request, the context, and the available actions.
- **Editable fields.** Mark specific fields as editable. Reviewers can rewrite content in the browser before acting. Edited values are returned to the workflow.
- **Configurable timeouts.** Set per request: auto-approve, auto-reject, or escalate to a secondary reviewer when the timeout expires. Workflows never hang indefinitely.
- **Reliable delivery.** Idempotent delivery queue — no duplicate sends, no webhook loops.
- **Full audit log.** Every decision (who acted, when, what was edited) is logged and queryable.

## Who It Is For

n8n builders running AI agent workflows that touch real customers, records, or money. Typical use cases include:

- AI-drafted emails that need a human to review before sending
- CRM record updates proposed by an AI agent
- Invoice or payment actions requiring manager approval
- Customer-facing messages that need sign-off before delivery
- Any workflow step where an unchecked AI action would be costly

Reviewers are often non-technical managers, clients, or teammates. They need a Humangent account but no n8n login or Slack seat.

## Technical Notes

- Integration: Standard HTTP Request node in n8n (POST to Humangent API)
- Callback: Humangent fires a POST to your n8n webhook URL when a decision is made
- Timeout: Configurable per request (seconds/minutes/hours)
- Timeout actions: auto-approve | auto-reject | escalate
- Custom actions: defined in the request payload, each maps to a different callback branch
- Editable fields: specified in the request payload, returned in the callback
- Auth: API key per workspace
- Audit log: queryable via API and visible in dashboard
- Humangent must be explicitly added to the workflow — it does not automatically intercept any agent calls

## Positioning

- **Category:** Human-in-the-loop approval platform for n8n
- **Target audience:** n8n builders running AI agent workflows
- **Core value:** Managed human decisions with visibility, control, and flexibility — not just a notification relay
- **Stage:** Pre-launch, private beta launching soon

## Founder

Iiro Rahkonen

## Pages

- / — Homepage and waitlist signup
- /privacy — Privacy Policy
- /terms — Terms of Service

## Waitlist

n8n builders can join the waitlist at https://humangent.io. The form asks for an email address and optionally what you are building in n8n.

## Guides (Blog)

In-depth guides on human-in-the-loop patterns in n8n. Each post is practitioner-led and grounded in real workflow patterns.

- [Human-in-the-Loop for n8n Workflows: Complete 2026 Guide](https://humangent.io/blog/n8n-human-in-the-loop-guide) — A practical guide to human-in-the-loop in n8n. Built-in features, DIY patterns, and external platforms, with a framework for picking the right one.
- [How to Build an n8n Approval Workflow Without Slack Duct-Tape](https://humangent.io/blog/n8n-approval-workflow-tutorial) — Three ways to build an n8n approval workflow: native Send-and-Wait, Wait-for-Webhook DIY, and a managed human-in-the-loop inbox. Node configs and trade-offs.
- [n8n AI Email Approval: Review Drafts Before Sending](https://humangent.io/blog/n8n-ai-email-approval-before-sending) — AI email agents hallucinate facts and miss tone. Build an n8n approval workflow where a human reviews, edits, or rejects every draft before it sends.
- [n8n CRM Approval: Stop AI From Overwriting Customer Data](https://humangent.io/blog/n8n-crm-approval-workflow) — Add an n8n CRM approval workflow so AI agents propose changes and humans review before anything hits your pipeline. Protect deal data at scale.
- [Why Every n8n AI Agent Workflow Needs Human Oversight](https://humangent.io/blog/why-n8n-ai-agents-need-human-oversight) — n8n AI agents act across dozens of tools. Without human approval at key moments, one mistake reaches customers. Here's where to draw the line.
- [n8n Approval Timeouts and Escalation: A Practical Guide](https://humangent.io/blog/n8n-approval-timeout-escalation) — How to handle approval timeouts and escalation in n8n. Compare built-in options, build multi-step chains, and stop workflows from hanging silently.
- [n8n Guardrails vs Human-in-the-Loop: When to Use Each](https://humangent.io/blog/n8n-guardrails-vs-hitl) — n8n guardrails catch rule violations automatically. HITL catches judgment calls. When to use each, and why production workflows need both.
- [n8n Social Media Approval: Review AI Posts Before Publishing](https://humangent.io/blog/n8n-ai-content-review-before-publishing) — AI writes social posts fast, but publishing unreviewed risks your brand. The n8n content approval workflow that adds human review before posts go live.
- [How to Build Multi-Level Approval Workflows in n8n](https://humangent.io/blog/n8n-multi-level-approval-workflow) — Three patterns for n8n multi-level approval workflows: sequential Send-and-Wait chains, database-driven polling, and a managed platform approach.
- [Audit Trails for n8n AI Agents: A Compliance Guide](https://humangent.io/blog/n8n-ai-agent-audit-trail-compliance) — n8n execution logs show a workflow ran. They don't show who approved what, when, or what they changed. How to build a real audit trail for AI agents.
- [n8n Payment and Refund Approval Workflows: Best Practices](https://humangent.io/blog/n8n-payment-refund-approval-workflow) — n8n financial workflows need strict human-in-the-loop controls. Refund and payment approval patterns with amount routing, dual approval, and audit.