Documentation

Advanced Search Settings

View as Markdown

Clear Ideas provides configurable search features at both the account and site level. Enable full‑text search, OCR for PDFs, and AI‑enhanced search to improve discovery while keeping your content private.

Account Settings

From Settings > Search:

Full‑Text Search: Index file and folder names for faster lookup.

OCR for PDFs: Extract text from PDFs so content becomes searchable and available to AI features.

AI Enhanced Search: Improves ranking and recall using AI embeddings. Your private data remains encrypted and is shared with models only as needed to fulfill queries.

Site‑Level Overrides

From a Site > Settings > Search you can override the account defaults:

  • Toggle full‑text and OCR settings for a specific site.
  • Enable AI Enhanced Search per site to tailor relevance for that site's audience.

How Site Settings Override Account Settings

Restriction Only: Site settings can only restrict features enabled at the account level. If an account-level feature is disabled, site settings cannot enable it.

Example Scenarios:

Account Enables, Site Disables:

  • Account Setting: fullTextSearchEnabled: true
  • Site Setting: Can set to false (disable for this site)
  • Result: ✅ Allowed - Site restricts account-level feature

Account Disables, Site Tries to Enable:

  • Account Setting: fullTextSearchEnabled: false
  • Site Setting: Cannot set to true (enable for this site)
  • Result: ❌ Not Allowed - Site cannot enable what account disables

Organization Policy Impact:

  • If organization policies control search settings, both account and site settings must comply
  • Policy indicators show when organization policies affect settings
  • See Organization Policies Detailed for more information

When to Use Site-Level Overrides

Site-Specific Needs: Use site-level overrides when different sites have different search requirements

Content Types: Configure search based on site content types (e.g., disable OCR for image-only sites)

Performance Optimization: Disable search features on sites where they're not needed to optimize performance

Collaboration Control: Control search capabilities per collaboration space based on team needs

Site Search Configuration

Navigate to Site Settings > Search to configure:

Site Search Preferences: Configure search settings specific to this site

Settings Available:

  • Full-text search toggle (if enabled at account level)
  • OCR for PDFs toggle (if enabled at account level)
  • AI Enhanced Search (if enabled at account level)

Policy Indicators: Visual indicators show when organization or account policies affect site settings

Changes to search settings may take time to apply as content (re)indexing completes.

Search supports structured metadata tokens using @key:value syntax.

Metadata Search becomes especially powerful when used with Extraction Workflows. Extraction workflows can classify uploaded files, extract structured fields, and save those fields back to the file so they can be searched later.

Examples:

  • @expiryDate:2026
  • @fullName:"Jane Doe"
  • @tag:invoice
  • @tags:invoice

You can combine plain text and metadata:

  • passport @expiryDate:2026
  • nda @tag:legal

In mixed queries, plain text drives retrieval (full-text or hybrid), and metadata filters are applied to narrow results.

Metadata-only queries are also supported. This is useful when you want to search by structured fields alone:

  • @vendorName:"Acme Corp"
  • @invoiceNumber:INV-1042
  • @tags:priority

Metadata Search and Extraction Workflows

When an Extraction Workflow runs successfully, its structured output is stored with the file metadata and becomes available for metadata search. This lets teams move from "find documents that mention a topic" to "find documents with a specific extracted field value."

Common examples include:

  • invoices by vendor, invoice number, due date, or amount
  • contracts by renewal date, governing law, or counterparty
  • ID documents by expiry date, full legal name, or document number
  • internal document routing by tags added during upload or review

Typical workflow:

  1. Configure an Extraction Workflow for a site or folder.
  2. Map upload variables and ensure the workflow returns structured output.
  3. Upload documents into the configured location.
  4. Use Metadata Search to query the extracted fields.

For setup details, see Extraction Workflows.

Built-in Fields

In addition to extracted metadata fields, search supports common built-in fields:

  • @contentType:image/png
  • @createdAt:2026
  • @updatedAt:2026
  • @name:passport
  • @tag:invoice
  • @tags:invoice
  • @size:1024
  • @extension:pdf
  • @folderId:<id>

@tag:<value> and @tags:<value> are equivalent. Both match content tags stored on the item.

Use quotes for values with spaces:

  • @fullName:"Jane Doe"

Practical Tips

  • Use extracted metadata fields when you need precise filtering.
  • Use plain text plus metadata together when you want relevance ranking plus narrowing.
  • Use @tag or @tags for manually assigned content tags.
  • If a field comes from an Extraction Workflow, confirm the workflow output is being saved consistently before relying on it in search.

Privacy and Security

All private data remains encrypted at rest. When AI features are enabled, only the minimum required content is decrypted and transmitted over encrypted channels to the selected model. See Encryption & Privacy.