Overview
Markloop is a web application designed to close the feedback loop between AI coding agents and human reviewers. It allows users to upload HTML documents generated by agents like Claude Code or Codex, share them with reviewers who can leave anchored comments and answer embedded questions, and then export the structured feedback back to the agent for automated application. The tool emphasizes preserving document formatting and providing context-rich feedback packages.
Key Features
- Anchored Comments: Reviewers can pin comments to specific sections or sentences within the document. Comments remain attached to the exact version they were made on, ensuring context is preserved across revisions.
- Agent-Ready Feedback Package: Every comment is exported with its CSS selector, quoted text, reviewer intent, surrounding context, and version number. This structured data can be consumed by AI agents via MCP (Model Context Protocol) or as plain markdown files.
- Version Tracking with Resolution Status: Each uploaded document becomes a version in a chain. Comments are marked as addressed or open, and new versions can track which feedback has been resolved. This allows agents to apply changes locally and publish updated versions.
- Unlimited Free Reviewers: Inviting teammates, clients, or stakeholders as reviewers incurs no additional cost. Reviewers see only the rendered document, not the source HTML, and cannot edit or access other projects.
Target Audience
Markloop is built for developers, product managers, consultants, and agencies who use AI coding agents to generate HTML documents such as specs, reports, proposals, and technical docs, and need a structured way to collect human feedback before finalizing.
