Settings

Settings is the connector's own configuration tab, and its subtitle draws the boundary: "Configure connector name, security, and cost controls." Where Connector Policy sets account-wide defaults, this tab overrides them for one connector.

R

Settings

Configure connector name, security, and cost controls.

General Settings

Server Name

DescriptionIssues, pull requests and code search for our repositories.

This helps AI agents understand when to use them.

DLP Protection

Prevent sensitive data from being sent to the AI client.

4 protected · glob-style matching (e.g. *.email, *.password)

*.email*.password*.ssn*.credit_card

Add pattern

Tool grouping override

Use global settingsFlat (no grouping)Grouped by tagAuto (threshold)

FinOps Guard

Set response limits and compression for cost control.

Max array items

50 items

Toon Compression

Compress tool descriptions to reduce token overhead from schema definitions.

Save

General settings

Server Name, the connector's display name, and its Description, with the help line that explains why it matters: "This helps AI agents understand when to use them." The description is not documentation for humans; it is part of what agents read when deciding whether this connector fits the task.

DLP Protection, per connector

The same Zero-Trust Data Loss Prevention the workspace sets globally, scoped here. Protected Fields are glob-style patterns (".email, .password", and deeper paths like items[*].credit_card), each removable, new ones added with the *.pattern input, and the count of protected fields shown live. This is the connector-level version of DLP Protection.

Tool grouping override

How the connector's endpoints become AI tools: Flat ("Each endpoint becomes an individual AI tool. Best for small APIs"), Grouped by tag ("Related endpoints share one AI tool with an action discriminator. Reduces token usage") or Auto, which groups automatically past a threshold. Use global settings defers to the account-wide Connector Settings choice. The AI Capabilities tab shows the result: this setting shapes it.

FinOps Guard, per connector

Response limits and compression before requests reach your budget: the Max array items slider truncates large array responses, and Toon Compression "Compress tool descriptions to reduce token overhead from schema definitions." The same controls exist account-wide in FinOps Guard; this is the per-connector override.

Why this tab exists

Every connector is different: an internal ERP may need its own redaction patterns, a huge API needs grouping, a chatty tool needs truncation. This tab is where the global policy meets the individual exception, with Save and the green "Settings saved successfully." at the end.