Shared documents. For you and your agents.
- Sharing unit
- A single doc. Every doc has a share link.
- Agent access
- First-class. MCP server, one tool call to publish, share, or edit.
A rich-text editor designed around humans co-editing with cursors and comments. Markland is the MCP-native shared surface for agents and humans as equal editors — here is how they compare.
Cursors, comments, suggest-mode — every affordance assumes a human at a keyboard. There is no MCP surface and no first-class publish, search, or iterate primitives for agents. When Google adds AI features, the agent is a guest in a human workflow. In Markland the agent is a first-class editor with its own token and the same toolset you use.
Cursors, comments, suggest-mode — every affordance assumes a human at a keyboard. There is no MCP surface and no first-class publish, search, or iterate primitives for agents. When Google adds AI features, the agent is a guest in a human workflow. In Markland the agent is a first-class editor with its own token and the same toolset you use.
Docs live inside a Drive UI your agent cannot navigate, organized in folders your agent cannot reason about. Markland docs are addressable by stable IDs, returned directly from every tool call, and searchable through `markland_search`.
Your agent produces markdown. Google Docs converts it into rich-text blocks, loses structure on export, and re-adds it on re-import. Markland stores the markdown the agent wrote and serves it back verbatim.
If the content is a meeting note with three humans leaving comments in real time, or a contract with a suggestion-mode review cycle, Google Docs is the right tool. Markland does not try to be that.
Teams of humans writing together, leaving comments, tracking suggestions, and using rich formatting.
Solo developers and small teams whose agents produce markdown they want to share — specs, plans, research notes, CLAUDE.md files.
Sign up, wire the MCP server into Claude Code, and skip the copy-paste.
Join the waitlist See all alternatives