# Ratifia

> Policy-driven human approval for AI workflows. Ratifia pauses an AI workflow for a human decision, presents the reviewer with the full AI context, and resumes with the verdict. It runs on top of an existing workflow engine and never executes your code.

## Docs — Start here

- [Core concepts](https://docs.ratifia.com/concepts): The decision object, callers, the inbox, and surfaces.
- [Quickstart](https://docs.ratifia.com/quickstart): Shortest path to a first decision.
- [Quickstart from code](https://docs.ratifia.com/quickstart/code): Two HTTP calls, no SDK — works from any language.
- [Glossary](https://docs.ratifia.com/glossary): Terms used throughout the docs.

## Docs — Integrating

- [MCP server](https://docs.ratifia.com/mcp): For agents that block on a verdict.
- [n8n](https://docs.ratifia.com/n8n): Approval inside an n8n workflow.
- [Enforce approval](https://docs.ratifia.com/enforce-approval): Verify and consume, so an approval cannot be reused or applied to a changed action.
- [API reference](https://docs.ratifia.com/api-reference): Every endpoint.

## Docs — Governing

- [Policies](https://docs.ratifia.com/policies): Who approves, quorum thresholds, and the surface.
- [Human in the loop](https://docs.ratifia.com/hitl): The approval model end to end.
- [RBAC](https://docs.ratifia.com/rbac): Roles and permissions.
- [Conversational approvals](https://docs.ratifia.com/conversational-approvals): Revising a proposed action before approving it.

## Docs — Surfaces

- [Notifications](https://docs.ratifia.com/notifications): How an approver is reached.
- [Slack](https://docs.ratifia.com/slack): Approve from a Slack message.
- [Discord](https://docs.ratifia.com/discord): Approve from a Discord message.
- [DocuSign](https://docs.ratifia.com/docusign): Signature-bound approval.

## Machine-readable

- [OpenAPI specification](https://docs.ratifia.com/openapi.json): the full HTTP API.

## Blog

- [Automation bias is a design bug](https://ratifia.com/blog/automation-bias-is-a-design-bug.md): Show an AI's recommendation the wrong way and your reviewers stop reviewing. Automation bias is not a training problem or a user problem. It is a layout decision.
- [The gate is a policy, not an if](https://ratifia.com/blog/the-gate-is-a-policy-not-an-if.md): There are two obvious ways to decide when a human is required, and both are traps: hard-code it and it rots; let the agent decide and it won't. The floor is the third way.
- [How did this come about?](https://ratifia.com/blog/why-we-are-writing-this.md): A blog about the unglamorous half of AI engineering: what happens when an agent proposes an action that a person has to stand behind.
- [What belongs in an approval context payload](https://ratifia.com/blog/what-belongs-in-an-approval-context-payload.md): A reviewer who approves without understanding the decision is a liability. Five things every approval request should carry — and three it never should.
- [Why waitForEvent is not human-in-the-loop](https://ratifia.com/blog/why-wait-for-event-is-not-human-in-the-loop.md): Every workflow engine can pause. Pausing is the easy half. The hard half is everything between pause and resume — the half you end up building yourself.

## Pages

- [Home](https://ratifia.com/): Product overview, the approval lifecycle, and decision surfaces.