Insight O' Mate

Advanced configuration

The Insight O' Mate desktop client can be configured using environment variables. These are primarily for power users, security teams, and those self-hosting the NLP engine.

General Config

  • IOM_LOG_ENGINE_REQUESTS Set to 1 to log every outbound request (the prompt and your field names) made to the NLP engine. Logs are appended to ~/.insight-o-mate/engine.log. Highly recommended for verifying privacy.
  • SIDECAR_CONFIG_DIR / SIDECAR_CONFIG_PATH Overrides the default directory (~/.insight-o-mate) where the app stores local configuration and saved connections.
  • INSIGHT_O_MATE_DISABLE_KEYCHAIN Set to 1 to disable using your OS's native secure keychain (macOS Keychain, Windows Credential Manager, libsecret). When disabled, secrets are stored in plaintext in your config file. Use with caution.

Network & NLP Engine

  • NLP_ENGINE_URL Overrides the URL of the stateless NLP engine. Use this if you are an Enterprise customer running a dedicated or self-hosted instance in your own VPC.
  • NLP_ENGINE_TIMEOUT_MS Overrides the default 15-second (15,000ms) timeout when waiting for the NLP engine to analyze a prompt.
  • PORTAL_URL Overrides the URL of the Insight O' Mate portal (used for billing and API key verification).
  • PORTAL_TIMEOUT_MS Overrides the default timeout for communicating with the portal.

Database Limits

The local sidecar aggregates data locally for operations your database can't do natively. You can tune these memory/performance limits:

  • IMA_MAX_SCAN_ROWS The maximum number of rows the app will scan during local aggregations (default: 100,000).
  • IMA_MAX_JOIN_ROWS The maximum number of rows the app will pull into memory for a client-side join (default: 50,000).
  • SLOW_QUERY_THRESHOLD_MS The threshold before a query is flagged as "slow" in the UI (default: 1,000ms).
Last updated: Jul 30 2026