Help

Start by choosing one topic you already care about. You can build a private Bucket, follow public Buckets in Discover, or connect a Bucket to your AI tool as grounding context.

Private Buckets are the default. Publish only when a Bucket is useful enough for other people to inspect.

Maintained mode lets MiraDock read new sources and propose Bucket updates. Add a URL, paste text, or upload a file, then review the changes against the source material.

MiraDock exposes Bucket knowledge through an MCP server so agent tools can search, read, create, and update selected Buckets without manual copy and paste.

New Buckets are created from your AI tool with a write token. Start with the MCP write setup guide.

{ "mcpServers": { "miradock": { "url": "https://miradock.com/mcp" } } }

MCP-created Buckets are private unless you publish them. Public Buckets can appear in Discover, be followed through creator profiles, and be connected by people who can view them.

Follow public Bucket owners whose work you trust. Their updates feed your home screen, and their knowledge bases become easier to find when you want grounded context in your AI tool.

Public Bucket parts include MCP, /llm.txt, /context.json, and markdown options so visitors can use the Bucket as read-only grounding context in their own AI tools.

Use the Connect panel on a Bucket part to copy the right endpoint or configure an MCP-aware tool with a read token.

Can I start blank? Yes. Templates are optional.

Where should I go next? Browse public Buckets.