Markloop

About Markloop

Markloop solves a workflow gap that only appeared once coding agents started writing prose: your agent drafts a spec, report, proposal, or technical doc as an HTML artifact, and then the review of that artifact has nowhere sensible to live. Paste it into a Google Doc and the layout, code blocks, and structure fall apart. Collect feedback in Slack, email, or screenshots and nothing is anchored to the line it is about — and you become the person hand-translating every comment back into a prompt. Markloop takes the HTML file, versions it inside a project, and gives you a shareable link. Viewers see the rendered document, not the source, and can comment on exact parts of it and answer your questions in context — no agent required on their side, and they can comment but not edit. Then your coding agent pulls the whole thread back over native MCP as structured context carrying target, quote, intent, and version, applies what is clear, asks you about what is not, and pushes the next version with `markloop_put_version`. MCP support is native for Claude Code and Codex; other agents can consume the same feedback as plain files. The vendor is explicit about scope — it is for HTML documents, not live websites or shipped apps — and pitches directly at the Claude Code artifact limitation, where artifacts stay inside your claude.ai org and viewers cannot comment on them.

Key Features

Tags