# Use cases Step-by-step scenarios for annotating documents with AI agents over MCP: marking up documents, running review cycles with replies, reporting on existing annotations, and rendering pages the agent can see. Step-by-step scenarios for annotating and reviewing documents with AI agents over MCP — each page answers one real question completely, with the config, the exact prompt, and the expected result: * [How to annotate documents with AI agents using MCP]({{< ref "annotation/mcp/use-cases/annotate-documents-with-ai-agents.md" >}}) — start here: the annotate → verify → iterate loop, and which file the agent writes to. * [How to run a document review cycle with an AI agent]({{< ref "annotation/mcp/use-cases/review-workflow-with-replies.md" >}}) — threads, authors, and resolving comments — a full review round driven from chat. * [How to extract and report on document annotations with AI]({{< ref "annotation/mcp/use-cases/extract-and-report-annotations.md" >}}) — turn a marked-up document into a summary, a table, or an XML archive. * [How to let an AI agent see the document page it annotated]({{< ref "annotation/mcp/use-cases/visual-preview-for-agents.md" >}}) — inline PNG rendering — the agent checks its own work instead of asserting it. * [Running GroupDocs MCP servers on-premise: architecture and security model]({{< ref "annotation/mcp/use-cases/on-premise-document-annotation.md" >}}) — architecture and security model for regulated environments — the page to send your security reviewer. New to the server? [Quick start]({{< ref "annotation/mcp/getting-started/_index.md" >}}) first.