Back to Documentation

Content search and discovery

As a project grows, finding what you need matters as much as storing it. Sutram offers complementary ways to discover content — by tag, by knowledge (full-text in the Wiki), by mention, and by browsing. This guide shows which one to use in each situation.

Choose the right method

You know… Use
A tag (category, author, year, patient…) Tag search, in the Content tab
The subject or you have a question Full-text search in the Wiki
The slug of a [[slug]] mention Resolve the mention
Where the item sits in the structure Folder browsing

Search by tag (items and folders)

Tags are key-value metadata. Tag search is the structured way to filter your collection.

  1. In the Content tab, open tag search
  2. Enter a key (e.g., category) and, optionally, a value (e.g., contract)
  3. To refine, add multiple conditions — all of them must be satisfied (AND logic)

How matching works:

  • Key: exact match, case-insensitive (Category = category)
  • Value: substring match, case-insensitive (neuro finds Neurology)
  • No value: returns any item that has that key, regardless of the value

Results include items and folders that match; from them you navigate straight to the content.

Search knowledge in the Wiki (full-text)

When you don't know the tag but know the subject — or you have a question — the Wiki is the way. Wiki search is full-text over the content: it covers the node name, the curated synthesis, and the full text of mirrored documents (with stemming), and returns a highlighted snippet per result.

The Wiki discovery flow:

  1. Search starting from the question or the topic
  2. Open the most relevant node and read the synthesis and the full text
  3. Follow the [[slug]] mentions and the backlinks to gather the surrounding context
  4. Use the graph to see how the topics connect

See Wiki and Knowledge Graph for the complete concept.

Follow mentions and slugs

Within the text, references in the [[slug]] format are clickable links. You can:

  • Resolve a slug to reach the item it points to
  • Open the backlinks of an item ("Mentioned in") to see every place that cites it

This is how you discover content by following the connections, instead of searching by words.

Search in Governed Content

The Governed Content space has its own search and respects need-to-know: you only find the Classes and documents you have permission to see, according to your role and each document's state. See Roles, permissions, and need-to-know.

AI-powered discovery (MCP)

An assistant connected through the Sutram MCP Server brings all these paths together in a conversation: search items and folders by tag, run full-text search in the Wiki, list nodes by name or category, and resolve mentions. Examples:

  • "Find all documents tagged with category = contract and year = 2026."
  • "What do we have about the refund policy? Search the wiki and summarize with sources."
  • "Where does the mention [[refund-policy]] point to?"

The reference for these tools is in the Sutram MCP Server Guide.

Tips

  • Tags for the structured, Wiki for the semantic — filter by metadata when you know the fields; ask the Wiki when you know the subject.
  • Standardize your tags — consistent values (preferably via enum, in Records) make tag search reliable.
  • Start broad, refine later — search by a single key, check the volume, and add conditions to narrow down.
  • Follow the backlinks — what you're looking for is often one hop away from an item you've already found.

Frequently Asked Questions

Q: I searched by tag and got nothing. What happened?

A: Check that the key exists (search by the key alone, without a value, to test), that there's no typo, and remember that the value is a substring match and the key is exact (both ignore case).

Q: What's the difference between tag search and Wiki search?

A: Tag search filters by structured metadata (key-value) — ideal when you know the fields. Wiki search is full-text over the content — ideal when you know the subject, not the fields.

Q: Can I search by the content inside the documents?

A: Yes, through the Wiki: the full text of mirrored documents is indexed, so full-text search reaches the body of the document, not just the title.

Next Steps

  • Wiki and Knowledge Graph — the knowledge-based discovery layer
  • Records and Metadata — how to structure the tags and fields that search uses
  • Sutram MCP Server Guide — the reference for the search tools for use with AI

Document Version: 1.0 Last Updated: July 2026 Author: Sutram Development Team

Back to Documentation