Skip to main content

​
What are Filtering Rules?

Filtering rules let you define precisely which files a connector should anchor. Without rules, a connector anchors every new file detected in the storage. Rules allow you to restrict anchoring to relevant files only — for example, only PDFs, or everything except temporary files. Rules are configured per connector and evaluated each time a new file is detected.

​
Default Behavior

If no filtering rules are configured on a connector, all files are anchored without restriction.

​
Rule Structure

Each rule has three parts:
PartDescription
ActionWhat to do when the rule matches: INCLUDE (anchor the file) or EXCLUDE (skip the file).
OperatorHow to combine multiple conditions within the rule: AND (all conditions must match) or OR (any condition must match).
ConditionsOne or more conditions that the file must satisfy.

​
Conditions

Each condition targets one field of the file and applies an operator to a value.
FieldAvailable OperatorsExample ValueDescription
extensionequalspdfFile extension, without the leading dot. Case-insensitive.
prefixstarts_with, equalsdocuments/reports/The file path or key prefix (useful for S3 prefixes or folder paths).
suffixends_with, equals_finalA suffix in the filename (excluding extension).
sizeless_than, greater_than1048576File size in bytes.
filename_containscontainsinvoiceA substring present anywhere in the filename (excluding path).

​
How Rules Are Evaluated

  1. When a new file is detected, all rules on the connector are evaluated in order.
  2. The first matching rule determines the outcome (INCLUDE or EXCLUDE).
  3. If no rule matches, the file is anchored by default (same as if no rules were configured).
Rules are evaluated top to bottom. Order matters when you have overlapping conditions. Place more specific rules before more general ones.

​
Examples

​
Anchor only PDFs

Action: INCLUDE
Operator: AND
Conditions:
  - extension equals pdf
Any file that is not a PDF will not be anchored, because no rule matches it and the default action applies — unless you add an explicit EXCLUDE rule for all other files. To strictly anchor only PDFs, add: 
Rule 1:
  Action: INCLUDE
  Operator: AND
  Conditions:
    - extension equals pdf

Rule 2:
  Action: EXCLUDE
  Operator: OR
  Conditions:
    - extension equals docx
    - extension equals xlsx
    - extension equals png
    ... (or handle via a more general rule)
Alternatively, structure it as a single INCLUDE rule and rely on the default-pass-through not applying since no INCLUDE rule matches other file types.

​
Exclude temporary files

Exclude files whose name contains common temporary indicators: 
Action: EXCLUDE
Operator: OR
Conditions:
  - filename_contains contains .tmp
  - filename_contains contains ~$
  - suffix ends_with _draft

​
Include only files above 1 MB

Action: INCLUDE
Operator: AND
Conditions:
  - size greater_than 1048576

​
Anchor invoices from a specific folder (S3)

Action: INCLUDE
Operator: AND
Conditions:
  - prefix starts_with invoices/2024/
  - extension equals pdf

​
Exclude large video files

Action: EXCLUDE
Operator: OR
Conditions:
  - extension equals mp4
  - extension equals mov
  - extension equals avi
  - size greater_than 524288000

​
Configuring Rules in the Dashboard

Rules are managed on the connector detail page in app.rootkey.ai:
  1. Open the connector.
  2. Go to the Rules tab.
  3. Click Add Rule to create a new rule.
  4. Set the action, operator, and add one or more conditions.
  5. Save. Rules take effect immediately for new files.
Changes to rules do not retroactively affect files that were already processed. Only new files detected after the rule change are evaluated against the updated rule set.
→ Back to Connectors Overview