Search

MDflow searches across document titles and markdown body text.

What you get in results

• Matching text snippets, shown directly in the result list, so you can tell which hit you want without opening each one.
• The containing folder for every result.
• Highlighted matching text.
• Title matches ranked first, on the assumption that a document named after your query is more likely what you meant than one that mentions it in passing.

Click a result to jump straight to the document.

Scope

Search is scoped to the active workspace. It searches everything you have in that workspace, and nothing from the others. Switch workspaces to search elsewhere.

This differs from the API and MCP server, which are account-wide and can see every workspace. See 01 Workspaces.

Encrypted documents are excluded

Encrypted documents never appear in search results.

This is not an oversight or a limitation to be fixed — it is the encryption guarantee working correctly. The server stores only ciphertext, so there is nothing readable to index. Nobody can search inside your encrypted documents, including MDflow.

The trade-off is real and worth deciding deliberately: encrypt what is genuinely sensitive, and accept that you will need to remember where it is. See 01 Document Encryption.

What search does not cover

• GitHub and Google Drive files. Those are read live from their own services and are never copied into your workspace, so they sit outside MDflow's search. Google Drive has its own search box within its sidebar section; GitHub does not.
• Comments.
• Version history.

Searching a shared collection

Readers of a collection shared in Knowledgebase mode (Pro) get their own full-text search across that collection's documents. It is language-neutral, diacritics-insensitive, and prefix-matching — typing private finds privately, and resume finds résumé.

Encrypted documents are not searchable there either, for the same reason.

See 03 Sharing Collections.

Search for agents

An AI agent does not use this search box. It calls mdflow_get_context with a topic, which ranks folder descriptions first, then folder names and document titles, and returns the most relevant markdown bodies.

This is why folder descriptions matter so much for agent workflows — see 02 Folders and Subfolders. Agents can also navigate precisely with mdflow_list_folders, mdflow_list_documents and mdflow_get_document. See 03 MCP Servers.